{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:F2TSDDAHZY4IOEA3TZ77KQMIDF","short_pith_number":"pith:F2TSDDAH","canonical_record":{"source":{"id":"2411.01077","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-11-01T23:18:32Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f260a360b5da63de6e30f8eb538a7acca170919b968b1a1db0d3547a0c0c9505","abstract_canon_sha256":"6ad4ac92199567556279e0baa4b4f44423cf31dd9f72403eee2c1fa3340002c2"},"schema_version":"1.0"},"canonical_sha256":"2ea7218c07ce3887101b9e7ff54188195c21d003dade77d0e51b82e21508bed9","source":{"kind":"arxiv","id":"2411.01077","version":5},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2411.01077","created_at":"2026-07-05T11:54:42Z"},{"alias_kind":"arxiv_version","alias_value":"2411.01077v5","created_at":"2026-07-05T11:54:42Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.01077","created_at":"2026-07-05T11:54:42Z"},{"alias_kind":"pith_short_12","alias_value":"F2TSDDAHZY4I","created_at":"2026-07-05T11:54:42Z"},{"alias_kind":"pith_short_16","alias_value":"F2TSDDAHZY4IOEA3","created_at":"2026-07-05T11:54:42Z"},{"alias_kind":"pith_short_8","alias_value":"F2TSDDAH","created_at":"2026-07-05T11:54:42Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:F2TSDDAHZY4IOEA3TZ77KQMIDF","target":"record","payload":{"canonical_record":{"source":{"id":"2411.01077","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-11-01T23:18:32Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"f260a360b5da63de6e30f8eb538a7acca170919b968b1a1db0d3547a0c0c9505","abstract_canon_sha256":"6ad4ac92199567556279e0baa4b4f44423cf31dd9f72403eee2c1fa3340002c2"},"schema_version":"1.0"},"canonical_sha256":"2ea7218c07ce3887101b9e7ff54188195c21d003dade77d0e51b82e21508bed9","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:54:42.448286Z","signature_b64":"St9EJ/PfH1uCK9rKQ+0P8q/hiP3CqgMpavDS0JPjnK9nFcxqUoW2jbTEJ3IFLfJFD7knX8gS3wpu7ZeifbCRDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2ea7218c07ce3887101b9e7ff54188195c21d003dade77d0e51b82e21508bed9","last_reissued_at":"2026-07-05T11:54:42.447800Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:54:42.447800Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2411.01077","source_version":5,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:54:42Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MCt9ksHY5SZVrLSY1aJNoLWqwSL7a+7BP+6l7TYIoJDjnB6SDHCVdZsKfUno7/2FevzFQdp9DfEI9w8cSccCAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T11:54:15.973632Z"},"content_sha256":"b430e28856dfb5cfbfa572caf96b7fdd64288c1aecbb8c8f3454db3ceedb1423","schema_version":"1.0","event_id":"sha256:b430e28856dfb5cfbfa572caf96b7fdd64288c1aecbb8c8f3454db3ceedb1423"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:F2TSDDAHZY4IOEA3TZ77KQMIDF","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Emoji Attack: Enhancing Jailbreak Attacks Against Judge LLM Detection","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"N. Benjamin Erichson, Yuqi Liu, Zhipeng Wei","submitted_at":"2024-11-01T23:18:32Z","abstract_excerpt":"Jailbreaking techniques trick Large Language Models (LLMs) into producing restricted output, posing a potential threat. One line of defense is to use another LLM as a Judge to evaluate the harmfulness of generated text. However, we reveal that these Judge LLMs are vulnerable to token segmentation bias, an issue that arises when delimiters alter the tokenization process, splitting words into smaller sub-tokens. This alters the embeddings of the entire sequence, reducing detection accuracy and allowing harmful content to be misclassified as safe. In this paper, we introduce Emoji Attack, a novel"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.01077","kind":"arxiv","version":5},"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/2411.01077/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":"2026-07-05T11:54:42Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"YtI7+FPWyviQVDio9saH3gS+Vz7Gt2QmczP3Ongmp8STvcwrVa89J0oJ4q/9MixAHDv6hB/Ux2nmXbq6EsDEBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T11:54:15.974134Z"},"content_sha256":"a90cc6121d778be45ffb99c8f96869305107628426f48ff72bc9e86cfd407075","schema_version":"1.0","event_id":"sha256:a90cc6121d778be45ffb99c8f96869305107628426f48ff72bc9e86cfd407075"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/F2TSDDAHZY4IOEA3TZ77KQMIDF/bundle.json","state_url":"https://pith.science/pith/F2TSDDAHZY4IOEA3TZ77KQMIDF/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/F2TSDDAHZY4IOEA3TZ77KQMIDF/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":"2026-08-04T11:54:15Z","links":{"resolver":"https://pith.science/pith/F2TSDDAHZY4IOEA3TZ77KQMIDF","bundle":"https://pith.science/pith/F2TSDDAHZY4IOEA3TZ77KQMIDF/bundle.json","state":"https://pith.science/pith/F2TSDDAHZY4IOEA3TZ77KQMIDF/state.json","well_known_bundle":"https://pith.science/.well-known/pith/F2TSDDAHZY4IOEA3TZ77KQMIDF/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:F2TSDDAHZY4IOEA3TZ77KQMIDF","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":"6ad4ac92199567556279e0baa4b4f44423cf31dd9f72403eee2c1fa3340002c2","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-11-01T23:18:32Z","title_canon_sha256":"f260a360b5da63de6e30f8eb538a7acca170919b968b1a1db0d3547a0c0c9505"},"schema_version":"1.0","source":{"id":"2411.01077","kind":"arxiv","version":5}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2411.01077","created_at":"2026-07-05T11:54:42Z"},{"alias_kind":"arxiv_version","alias_value":"2411.01077v5","created_at":"2026-07-05T11:54:42Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.01077","created_at":"2026-07-05T11:54:42Z"},{"alias_kind":"pith_short_12","alias_value":"F2TSDDAHZY4I","created_at":"2026-07-05T11:54:42Z"},{"alias_kind":"pith_short_16","alias_value":"F2TSDDAHZY4IOEA3","created_at":"2026-07-05T11:54:42Z"},{"alias_kind":"pith_short_8","alias_value":"F2TSDDAH","created_at":"2026-07-05T11:54:42Z"}],"graph_snapshots":[{"event_id":"sha256:a90cc6121d778be45ffb99c8f96869305107628426f48ff72bc9e86cfd407075","target":"graph","created_at":"2026-07-05T11:54:42Z","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/2411.01077/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Jailbreaking techniques trick Large Language Models (LLMs) into producing restricted output, posing a potential threat. One line of defense is to use another LLM as a Judge to evaluate the harmfulness of generated text. However, we reveal that these Judge LLMs are vulnerable to token segmentation bias, an issue that arises when delimiters alter the tokenization process, splitting words into smaller sub-tokens. This alters the embeddings of the entire sequence, reducing detection accuracy and allowing harmful content to be misclassified as safe. In this paper, we introduce Emoji Attack, a novel","authors_text":"N. Benjamin Erichson, Yuqi Liu, Zhipeng Wei","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-11-01T23:18:32Z","title":"Emoji Attack: Enhancing Jailbreak Attacks Against Judge LLM Detection"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.01077","kind":"arxiv","version":5},"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:b430e28856dfb5cfbfa572caf96b7fdd64288c1aecbb8c8f3454db3ceedb1423","target":"record","created_at":"2026-07-05T11:54:42Z","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":"6ad4ac92199567556279e0baa4b4f44423cf31dd9f72403eee2c1fa3340002c2","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-11-01T23:18:32Z","title_canon_sha256":"f260a360b5da63de6e30f8eb538a7acca170919b968b1a1db0d3547a0c0c9505"},"schema_version":"1.0","source":{"id":"2411.01077","kind":"arxiv","version":5}},"canonical_sha256":"2ea7218c07ce3887101b9e7ff54188195c21d003dade77d0e51b82e21508bed9","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2ea7218c07ce3887101b9e7ff54188195c21d003dade77d0e51b82e21508bed9","first_computed_at":"2026-07-05T11:54:42.447800Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:54:42.447800Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"St9EJ/PfH1uCK9rKQ+0P8q/hiP3CqgMpavDS0JPjnK9nFcxqUoW2jbTEJ3IFLfJFD7knX8gS3wpu7ZeifbCRDg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:54:42.448286Z","signed_message":"canonical_sha256_bytes"},"source_id":"2411.01077","source_kind":"arxiv","source_version":5}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b430e28856dfb5cfbfa572caf96b7fdd64288c1aecbb8c8f3454db3ceedb1423","sha256:a90cc6121d778be45ffb99c8f96869305107628426f48ff72bc9e86cfd407075"],"state_sha256":"1d023b2956d079ab69542734a9cac14bb8b4c6afd79e70395c2b36619c289c1d"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+lGcwZcm//XozTcKhwBBHQb4eJXwpPFies96L6MH+ArIfAM1JdtBBRinZcUlFjAO5VmcCa9+uLkFINIIZ3flAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T11:54:15.978682Z","bundle_sha256":"fc11de2f4b6bb22adfc95d04692d6025a87bb0c769c1f9966a291ac3a11b4534"}}