{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:63CAIHLK2QGOWBWGWKH7R5TU2U","short_pith_number":"pith:63CAIHLK","canonical_record":{"source":{"id":"2404.09894","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-15T16:06:36Z","cross_cats_sorted":["cs.SE"],"title_canon_sha256":"ae6332078aa53618d407c2a1190c1b6c46e8eb66f7a8a29b09bc79faa2d53ccb","abstract_canon_sha256":"f3e723000f16f0bf1bc19bf2b278896d76865eaf49c509660d4cf7aafb58dfdc"},"schema_version":"1.0"},"canonical_sha256":"f6c4041d6ad40ceb06c6b28ff8f674d5370a5cdc0dceee9543167d7c1e1aa893","source":{"kind":"arxiv","id":"2404.09894","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.09894","created_at":"2026-07-05T08:09:50Z"},{"alias_kind":"arxiv_version","alias_value":"2404.09894v3","created_at":"2026-07-05T08:09:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.09894","created_at":"2026-07-05T08:09:50Z"},{"alias_kind":"pith_short_12","alias_value":"63CAIHLK2QGO","created_at":"2026-07-05T08:09:50Z"},{"alias_kind":"pith_short_16","alias_value":"63CAIHLK2QGOWBWG","created_at":"2026-07-05T08:09:50Z"},{"alias_kind":"pith_short_8","alias_value":"63CAIHLK","created_at":"2026-07-05T08:09:50Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:63CAIHLK2QGOWBWGWKH7R5TU2U","target":"record","payload":{"canonical_record":{"source":{"id":"2404.09894","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-15T16:06:36Z","cross_cats_sorted":["cs.SE"],"title_canon_sha256":"ae6332078aa53618d407c2a1190c1b6c46e8eb66f7a8a29b09bc79faa2d53ccb","abstract_canon_sha256":"f3e723000f16f0bf1bc19bf2b278896d76865eaf49c509660d4cf7aafb58dfdc"},"schema_version":"1.0"},"canonical_sha256":"f6c4041d6ad40ceb06c6b28ff8f674d5370a5cdc0dceee9543167d7c1e1aa893","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:09:50.127505Z","signature_b64":"rSHX/GaG9Emr+Y5OZOckWnrA5pdT+phVh6Csn0PewrhhFVlGs45ou+/yCCn+nvFTB7p6K1CjLriCbd9CAPa5BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f6c4041d6ad40ceb06c6b28ff8f674d5370a5cdc0dceee9543167d7c1e1aa893","last_reissued_at":"2026-07-05T08:09:50.127024Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:09:50.127024Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2404.09894","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":"2026-07-05T08:09:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"sYbsObD4/eCTc8M0lqq4/l7F6ENrl0mzYc08pHt7wlKCoC5Bixv9NatT9FBWKZFRzYYGWxVQBlTvRYnR6CmhCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T08:27:34.132690Z"},"content_sha256":"105226cf665e0489b95811584554ad3d1286ebaa3892c991b561cad3d6c676f8","schema_version":"1.0","event_id":"sha256:105226cf665e0489b95811584554ad3d1286ebaa3892c991b561cad3d6c676f8"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:63CAIHLK2QGOWBWGWKH7R5TU2U","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Glitch Tokens in Large Language Models: Categorization Taxonomy and Effective Detection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SE"],"primary_cat":"cs.CL","authors_text":"Gelei Deng, Haoyu Wang, Kailong Wang, Ling Shi, Wenjia Song, Yang Liu, Yi Liu, Ying Zhang, Yuekang Li, Yuxi Li","submitted_at":"2024-04-15T16:06:36Z","abstract_excerpt":"With the expanding application of Large Language Models (LLMs) in various domains, it becomes imperative to comprehensively investigate their unforeseen behaviors and consequent outcomes. In this study, we introduce and systematically explore the phenomenon of \"glitch tokens\", which are anomalous tokens produced by established tokenizers and could potentially compromise the models' quality of response. Specifically, we experiment on seven top popular LLMs utilizing three distinct tokenizers and involving a totally of 182,517 tokens. We present categorizations of the identified glitch tokens an"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.09894","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/2404.09894/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":"2026-07-05T08:09:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ei/RFOu1D6hWcQKjUxpmw4r03BLvGbKnU0WBu4IzCnX6zYqGkNmr/Skxcq2aDYNscKvmKBadd807/IVWNBlmDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T08:27:34.133651Z"},"content_sha256":"aa1b62e0885f6e8005fd288ee20decd522e212fdca15e4a193c54cac56c3ef57","schema_version":"1.0","event_id":"sha256:aa1b62e0885f6e8005fd288ee20decd522e212fdca15e4a193c54cac56c3ef57"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/63CAIHLK2QGOWBWGWKH7R5TU2U/bundle.json","state_url":"https://pith.science/pith/63CAIHLK2QGOWBWGWKH7R5TU2U/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/63CAIHLK2QGOWBWGWKH7R5TU2U/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":"2026-08-21T08:27:34Z","links":{"resolver":"https://pith.science/pith/63CAIHLK2QGOWBWGWKH7R5TU2U","bundle":"https://pith.science/pith/63CAIHLK2QGOWBWGWKH7R5TU2U/bundle.json","state":"https://pith.science/pith/63CAIHLK2QGOWBWGWKH7R5TU2U/state.json","well_known_bundle":"https://pith.science/.well-known/pith/63CAIHLK2QGOWBWGWKH7R5TU2U/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:63CAIHLK2QGOWBWGWKH7R5TU2U","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":"f3e723000f16f0bf1bc19bf2b278896d76865eaf49c509660d4cf7aafb58dfdc","cross_cats_sorted":["cs.SE"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-15T16:06:36Z","title_canon_sha256":"ae6332078aa53618d407c2a1190c1b6c46e8eb66f7a8a29b09bc79faa2d53ccb"},"schema_version":"1.0","source":{"id":"2404.09894","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.09894","created_at":"2026-07-05T08:09:50Z"},{"alias_kind":"arxiv_version","alias_value":"2404.09894v3","created_at":"2026-07-05T08:09:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.09894","created_at":"2026-07-05T08:09:50Z"},{"alias_kind":"pith_short_12","alias_value":"63CAIHLK2QGO","created_at":"2026-07-05T08:09:50Z"},{"alias_kind":"pith_short_16","alias_value":"63CAIHLK2QGOWBWG","created_at":"2026-07-05T08:09:50Z"},{"alias_kind":"pith_short_8","alias_value":"63CAIHLK","created_at":"2026-07-05T08:09:50Z"}],"graph_snapshots":[{"event_id":"sha256:aa1b62e0885f6e8005fd288ee20decd522e212fdca15e4a193c54cac56c3ef57","target":"graph","created_at":"2026-07-05T08:09:50Z","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/2404.09894/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"With the expanding application of Large Language Models (LLMs) in various domains, it becomes imperative to comprehensively investigate their unforeseen behaviors and consequent outcomes. In this study, we introduce and systematically explore the phenomenon of \"glitch tokens\", which are anomalous tokens produced by established tokenizers and could potentially compromise the models' quality of response. Specifically, we experiment on seven top popular LLMs utilizing three distinct tokenizers and involving a totally of 182,517 tokens. We present categorizations of the identified glitch tokens an","authors_text":"Gelei Deng, Haoyu Wang, Kailong Wang, Ling Shi, Wenjia Song, Yang Liu, Yi Liu, Ying Zhang, Yuekang Li, Yuxi Li","cross_cats":["cs.SE"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-15T16:06:36Z","title":"Glitch Tokens in Large Language Models: Categorization Taxonomy and Effective Detection"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.09894","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":"sha256:105226cf665e0489b95811584554ad3d1286ebaa3892c991b561cad3d6c676f8","target":"record","created_at":"2026-07-05T08:09:50Z","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":"f3e723000f16f0bf1bc19bf2b278896d76865eaf49c509660d4cf7aafb58dfdc","cross_cats_sorted":["cs.SE"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-04-15T16:06:36Z","title_canon_sha256":"ae6332078aa53618d407c2a1190c1b6c46e8eb66f7a8a29b09bc79faa2d53ccb"},"schema_version":"1.0","source":{"id":"2404.09894","kind":"arxiv","version":3}},"canonical_sha256":"f6c4041d6ad40ceb06c6b28ff8f674d5370a5cdc0dceee9543167d7c1e1aa893","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f6c4041d6ad40ceb06c6b28ff8f674d5370a5cdc0dceee9543167d7c1e1aa893","first_computed_at":"2026-07-05T08:09:50.127024Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:09:50.127024Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"rSHX/GaG9Emr+Y5OZOckWnrA5pdT+phVh6Csn0PewrhhFVlGs45ou+/yCCn+nvFTB7p6K1CjLriCbd9CAPa5BQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:09:50.127505Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.09894","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:105226cf665e0489b95811584554ad3d1286ebaa3892c991b561cad3d6c676f8","sha256:aa1b62e0885f6e8005fd288ee20decd522e212fdca15e4a193c54cac56c3ef57"],"state_sha256":"7b1ca23cf93f12453c856c433f58c812c6da942967f949c529ffa12ec5fe6407"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"LUAPzkyTDBEe1LYU8VscLCQ/YnGIcPHHQLCsn5TsROPRoFKi7oaYhcoo5YePgEWhqbw3JqEiim4AEyBHnwMMBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-21T08:27:34.139940Z","bundle_sha256":"1fed6e1aef90f2745d98d768305dd409b75dba46202c757856f9abd1443af058"}}