{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YEFWVZHO3DSJFCCXAG2EFWUAKL","short_pith_number":"pith:YEFWVZHO","schema_version":"1.0","canonical_sha256":"c10b6ae4eed8e492885701b442da8052d2a8be6dced57cf21fad7ff2cd9928b6","source":{"kind":"arxiv","id":"2506.05502","version":1},"attestation_state":"computed","paper":{"title":"StealthInk: A Multi-bit and Stealthy Watermark for Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Brian Mark, Chuxiong Wu, Kai Zeng, Massieh Kordi Boroujeny, Ya Jiang","submitted_at":"2025-06-05T18:37:38Z","abstract_excerpt":"Watermarking for large language models (LLMs) offers a promising approach to identifying AI-generated text. Existing approaches, however, either compromise the distribution of original generated text by LLMs or are limited to embedding zero-bit information that only allows for watermark detection but ignores identification. We present StealthInk, a stealthy multi-bit watermarking scheme that preserves the original text distribution while enabling the embedding of provenance data, such as userID, TimeStamp, and modelID, within LLM-generated text. This enhances fast traceability without requirin"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2506.05502","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-06-05T18:37:38Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"6d6259a30f7d452327f2db07b6121fd71ce759107d91f734bbcefbf9cf7dfd91","abstract_canon_sha256":"2ba9ef52097053cc40851f1b8045be5dcd1eb3250b9634b2950bd011d6063cc1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:17:03.931352Z","signature_b64":"pdMz3oAz5HGR3FyQePTHCxwEk4nK824Rr7MLKU1ibXUyic3FHPT2SDRdt1VHr+E/4iVAn4vIQMoQSHs9eMvsAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c10b6ae4eed8e492885701b442da8052d2a8be6dced57cf21fad7ff2cd9928b6","last_reissued_at":"2026-07-05T11:17:03.930886Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:17:03.930886Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"StealthInk: A Multi-bit and Stealthy Watermark for Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Brian Mark, Chuxiong Wu, Kai Zeng, Massieh Kordi Boroujeny, Ya Jiang","submitted_at":"2025-06-05T18:37:38Z","abstract_excerpt":"Watermarking for large language models (LLMs) offers a promising approach to identifying AI-generated text. Existing approaches, however, either compromise the distribution of original generated text by LLMs or are limited to embedding zero-bit information that only allows for watermark detection but ignores identification. We present StealthInk, a stealthy multi-bit watermarking scheme that preserves the original text distribution while enabling the embedding of provenance data, such as userID, TimeStamp, and modelID, within LLM-generated text. This enhances fast traceability without requirin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.05502","kind":"arxiv","version":1},"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/2506.05502/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2506.05502","created_at":"2026-07-05T11:17:03.930942+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.05502v1","created_at":"2026-07-05T11:17:03.930942+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.05502","created_at":"2026-07-05T11:17:03.930942+00:00"},{"alias_kind":"pith_short_12","alias_value":"YEFWVZHO3DSJ","created_at":"2026-07-05T11:17:03.930942+00:00"},{"alias_kind":"pith_short_16","alias_value":"YEFWVZHO3DSJFCCX","created_at":"2026-07-05T11:17:03.930942+00:00"},{"alias_kind":"pith_short_8","alias_value":"YEFWVZHO","created_at":"2026-07-05T11:17:03.930942+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.11653","citing_title":"Every Bit, Everywhere, All at Once: A Binomial Multibit LLM Watermark","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12216","citing_title":"TimeMark: A Trustworthy Time Watermarking Framework for Exact Generation-Time Recovery from AIGC","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08759","citing_title":"Optimal Multi-bit Generative Watermarking Schemes Under Worst-Case False-Alarm Constraints","ref_index":29,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YEFWVZHO3DSJFCCXAG2EFWUAKL","json":"https://pith.science/pith/YEFWVZHO3DSJFCCXAG2EFWUAKL.json","graph_json":"https://pith.science/api/pith-number/YEFWVZHO3DSJFCCXAG2EFWUAKL/graph.json","events_json":"https://pith.science/api/pith-number/YEFWVZHO3DSJFCCXAG2EFWUAKL/events.json","paper":"https://pith.science/paper/YEFWVZHO"},"agent_actions":{"view_html":"https://pith.science/pith/YEFWVZHO3DSJFCCXAG2EFWUAKL","download_json":"https://pith.science/pith/YEFWVZHO3DSJFCCXAG2EFWUAKL.json","view_paper":"https://pith.science/paper/YEFWVZHO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.05502&json=true","fetch_graph":"https://pith.science/api/pith-number/YEFWVZHO3DSJFCCXAG2EFWUAKL/graph.json","fetch_events":"https://pith.science/api/pith-number/YEFWVZHO3DSJFCCXAG2EFWUAKL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YEFWVZHO3DSJFCCXAG2EFWUAKL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YEFWVZHO3DSJFCCXAG2EFWUAKL/action/storage_attestation","attest_author":"https://pith.science/pith/YEFWVZHO3DSJFCCXAG2EFWUAKL/action/author_attestation","sign_citation":"https://pith.science/pith/YEFWVZHO3DSJFCCXAG2EFWUAKL/action/citation_signature","submit_replication":"https://pith.science/pith/YEFWVZHO3DSJFCCXAG2EFWUAKL/action/replication_record"}},"created_at":"2026-07-05T11:17:03.930942+00:00","updated_at":"2026-07-05T11:17:03.930942+00:00"}