{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:SMN6VH6TERO7QXX66PPTJSLI2Z","short_pith_number":"pith:SMN6VH6T","canonical_record":{"source":{"id":"2412.18171","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2024-12-24T05:10:02Z","cross_cats_sorted":[],"title_canon_sha256":"4d7f3ddf6887c0762053f6ac851534124c937470f2b27d5dd04201c6444135d0","abstract_canon_sha256":"7241f7a83b3f20bf2aaf14e6c46cbc6c0a21c12a181001fb2541bc0d5af3736a"},"schema_version":"1.0"},"canonical_sha256":"931bea9fd3245df85efef3df34c968d65dcaa6c88c766558e96e85ab87490e05","source":{"kind":"arxiv","id":"2412.18171","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2412.18171","created_at":"2026-07-05T09:54:07Z"},{"alias_kind":"arxiv_version","alias_value":"2412.18171v2","created_at":"2026-07-05T09:54:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.18171","created_at":"2026-07-05T09:54:07Z"},{"alias_kind":"pith_short_12","alias_value":"SMN6VH6TERO7","created_at":"2026-07-05T09:54:07Z"},{"alias_kind":"pith_short_16","alias_value":"SMN6VH6TERO7QXX6","created_at":"2026-07-05T09:54:07Z"},{"alias_kind":"pith_short_8","alias_value":"SMN6VH6T","created_at":"2026-07-05T09:54:07Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:SMN6VH6TERO7QXX66PPTJSLI2Z","target":"record","payload":{"canonical_record":{"source":{"id":"2412.18171","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2024-12-24T05:10:02Z","cross_cats_sorted":[],"title_canon_sha256":"4d7f3ddf6887c0762053f6ac851534124c937470f2b27d5dd04201c6444135d0","abstract_canon_sha256":"7241f7a83b3f20bf2aaf14e6c46cbc6c0a21c12a181001fb2541bc0d5af3736a"},"schema_version":"1.0"},"canonical_sha256":"931bea9fd3245df85efef3df34c968d65dcaa6c88c766558e96e85ab87490e05","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:54:07.777902Z","signature_b64":"rzoSC3blqMN+lSKdZITzYQr5Dr1jn0a46S3L9V1UKTgIhRj1xzjhhunGRiuF74C8+mhmUbZovBy8Z+KE46U3DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"931bea9fd3245df85efef3df34c968d65dcaa6c88c766558e96e85ab87490e05","last_reissued_at":"2026-07-05T09:54:07.777451Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:54:07.777451Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2412.18171","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":"2026-07-05T09:54:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"KbwKZewHDaZxm6I4BuiUzlwott7X8kyOYfMlGnHps6Jk1M8zUkZYOC7gmoXOy+SKBiv8DmcotOgGZKlytB4wBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-13T07:06:06.845468Z"},"content_sha256":"85ffe7f371f2b4ab4fdd00927825423efb029a04eaa1d1f06f8e5cdd46ef6df3","schema_version":"1.0","event_id":"sha256:85ffe7f371f2b4ab4fdd00927825423efb029a04eaa1d1f06f8e5cdd46ef6df3"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:SMN6VH6TERO7QXX66PPTJSLI2Z","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Token Highlighter: Inspecting and Mitigating Jailbreak Prompts for Large Language Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Pin-Yu Chen, Tsung-Yi Ho, Xiaomeng Hu","submitted_at":"2024-12-24T05:10:02Z","abstract_excerpt":"Large Language Models (LLMs) are increasingly being integrated into services such as ChatGPT to provide responses to user queries. To mitigate potential harm and prevent misuse, there have been concerted efforts to align the LLMs with human values and legal compliance by incorporating various techniques, such as Reinforcement Learning from Human Feedback (RLHF), into the training of the LLMs. However, recent research has exposed that even aligned LLMs are susceptible to adversarial manipulations known as Jailbreak Attacks. To address this challenge, this paper proposes a method called Token Hi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.18171","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/2412.18171/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":"2026-07-05T09:54:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qksI/GyL3gFlh6O7gnaBadNmBhuZ3G8S66C1YE7lNwpd5ppvEhzSj0j7+RtEt6maLqNofuOwHkdrVNnZ+kduBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-13T07:06:06.845866Z"},"content_sha256":"d8fe5f9960ecbe692c6aed30481c0f18ee976fd293006f2c7bd8d14e1c9d2bbe","schema_version":"1.0","event_id":"sha256:d8fe5f9960ecbe692c6aed30481c0f18ee976fd293006f2c7bd8d14e1c9d2bbe"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/SMN6VH6TERO7QXX66PPTJSLI2Z/bundle.json","state_url":"https://pith.science/pith/SMN6VH6TERO7QXX66PPTJSLI2Z/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/SMN6VH6TERO7QXX66PPTJSLI2Z/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":"2026-08-13T07:06:06Z","links":{"resolver":"https://pith.science/pith/SMN6VH6TERO7QXX66PPTJSLI2Z","bundle":"https://pith.science/pith/SMN6VH6TERO7QXX66PPTJSLI2Z/bundle.json","state":"https://pith.science/pith/SMN6VH6TERO7QXX66PPTJSLI2Z/state.json","well_known_bundle":"https://pith.science/.well-known/pith/SMN6VH6TERO7QXX66PPTJSLI2Z/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:SMN6VH6TERO7QXX66PPTJSLI2Z","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":"7241f7a83b3f20bf2aaf14e6c46cbc6c0a21c12a181001fb2541bc0d5af3736a","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2024-12-24T05:10:02Z","title_canon_sha256":"4d7f3ddf6887c0762053f6ac851534124c937470f2b27d5dd04201c6444135d0"},"schema_version":"1.0","source":{"id":"2412.18171","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2412.18171","created_at":"2026-07-05T09:54:07Z"},{"alias_kind":"arxiv_version","alias_value":"2412.18171v2","created_at":"2026-07-05T09:54:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.18171","created_at":"2026-07-05T09:54:07Z"},{"alias_kind":"pith_short_12","alias_value":"SMN6VH6TERO7","created_at":"2026-07-05T09:54:07Z"},{"alias_kind":"pith_short_16","alias_value":"SMN6VH6TERO7QXX6","created_at":"2026-07-05T09:54:07Z"},{"alias_kind":"pith_short_8","alias_value":"SMN6VH6T","created_at":"2026-07-05T09:54:07Z"}],"graph_snapshots":[{"event_id":"sha256:d8fe5f9960ecbe692c6aed30481c0f18ee976fd293006f2c7bd8d14e1c9d2bbe","target":"graph","created_at":"2026-07-05T09:54:07Z","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/2412.18171/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs) are increasingly being integrated into services such as ChatGPT to provide responses to user queries. To mitigate potential harm and prevent misuse, there have been concerted efforts to align the LLMs with human values and legal compliance by incorporating various techniques, such as Reinforcement Learning from Human Feedback (RLHF), into the training of the LLMs. However, recent research has exposed that even aligned LLMs are susceptible to adversarial manipulations known as Jailbreak Attacks. To address this challenge, this paper proposes a method called Token Hi","authors_text":"Pin-Yu Chen, Tsung-Yi Ho, Xiaomeng Hu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2024-12-24T05:10:02Z","title":"Token Highlighter: Inspecting and Mitigating Jailbreak Prompts for Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.18171","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":"sha256:85ffe7f371f2b4ab4fdd00927825423efb029a04eaa1d1f06f8e5cdd46ef6df3","target":"record","created_at":"2026-07-05T09:54:07Z","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":"7241f7a83b3f20bf2aaf14e6c46cbc6c0a21c12a181001fb2541bc0d5af3736a","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2024-12-24T05:10:02Z","title_canon_sha256":"4d7f3ddf6887c0762053f6ac851534124c937470f2b27d5dd04201c6444135d0"},"schema_version":"1.0","source":{"id":"2412.18171","kind":"arxiv","version":2}},"canonical_sha256":"931bea9fd3245df85efef3df34c968d65dcaa6c88c766558e96e85ab87490e05","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"931bea9fd3245df85efef3df34c968d65dcaa6c88c766558e96e85ab87490e05","first_computed_at":"2026-07-05T09:54:07.777451Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:54:07.777451Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"rzoSC3blqMN+lSKdZITzYQr5Dr1jn0a46S3L9V1UKTgIhRj1xzjhhunGRiuF74C8+mhmUbZovBy8Z+KE46U3DA==","signature_status":"signed_v1","signed_at":"2026-07-05T09:54:07.777902Z","signed_message":"canonical_sha256_bytes"},"source_id":"2412.18171","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:85ffe7f371f2b4ab4fdd00927825423efb029a04eaa1d1f06f8e5cdd46ef6df3","sha256:d8fe5f9960ecbe692c6aed30481c0f18ee976fd293006f2c7bd8d14e1c9d2bbe"],"state_sha256":"ad2957c9611b3551763c9c771490a0b037bdd829ed72a223055d683a6b5de35f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/HJ9y0X6R0cLKRpi+V33qRhjbR2kzk8LDA7qpcW8K1M3yTrbG1vIs6yqgPl1uPAEF8RPaScV9SAOL2e4HHamDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-13T07:06:06.848890Z","bundle_sha256":"2e31e57d1f9d629f96811452ad392c570008d3718fb2ac76eb60968d94cbe060"}}