{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:KNJVBENMTS3USTJJWC6GX5KK5F","short_pith_number":"pith:KNJVBENM","schema_version":"1.0","canonical_sha256":"53535091ac9cb7494d29b0bc6bf54ae947adcd0e9112376f83b77513ae4f8b1a","source":{"kind":"arxiv","id":"2401.16820","version":5},"attestation_state":"computed","paper":{"title":"Provably Robust Multi-bit Watermarking for AI-generated Text","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Dong Yin, Jiaheng Zhang, Jinyuan Jia, Tianyang Tao, Wei Zou, Wengrui Zheng, Wenjie Qu, Yanze Jiang, Zhihua Tian","submitted_at":"2024-01-30T08:46:48Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated remarkable capabilities of generating texts resembling human language. However, they can be misused by criminals to create deceptive content, such as fake news and phishing emails, which raises ethical concerns. Watermarking is a key technique to address these concerns, which embeds a message (e.g., a bit string) into a text generated by an LLM. By embedding the user ID (represented as a bit string) into generated texts, we can trace generated texts to the user, known as content source tracing. The major limitation of existing watermarking techniq"},"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":"2401.16820","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2024-01-30T08:46:48Z","cross_cats_sorted":[],"title_canon_sha256":"50f46f99988697b878ddde3928a5356305c8b3a2d2d57de2fc27bdd10d2f1793","abstract_canon_sha256":"71a5926721e735280bf8c4c86f9140c18de85c54ec0547db15e076af2d05ad81"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:06:02.966417Z","signature_b64":"SlYn7su66B+deBgFyq4xZTJ6GblasLjFUfjuzVHCo35thmSkVaNQIUazJNB8XOokBwM148IXNVt9Jh5yV9hSBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"53535091ac9cb7494d29b0bc6bf54ae947adcd0e9112376f83b77513ae4f8b1a","last_reissued_at":"2026-07-05T10:06:02.965935Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:06:02.965935Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Provably Robust Multi-bit Watermarking for AI-generated Text","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Dong Yin, Jiaheng Zhang, Jinyuan Jia, Tianyang Tao, Wei Zou, Wengrui Zheng, Wenjie Qu, Yanze Jiang, Zhihua Tian","submitted_at":"2024-01-30T08:46:48Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated remarkable capabilities of generating texts resembling human language. However, they can be misused by criminals to create deceptive content, such as fake news and phishing emails, which raises ethical concerns. Watermarking is a key technique to address these concerns, which embeds a message (e.g., a bit string) into a text generated by an LLM. By embedding the user ID (represented as a bit string) into generated texts, we can trace generated texts to the user, known as content source tracing. The major limitation of existing watermarking techniq"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.16820","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/2401.16820/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":"2401.16820","created_at":"2026-07-05T10:06:02.965989+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.16820v5","created_at":"2026-07-05T10:06:02.965989+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.16820","created_at":"2026-07-05T10:06:02.965989+00:00"},{"alias_kind":"pith_short_12","alias_value":"KNJVBENMTS3U","created_at":"2026-07-05T10:06:02.965989+00:00"},{"alias_kind":"pith_short_16","alias_value":"KNJVBENMTS3USTJJ","created_at":"2026-07-05T10:06:02.965989+00:00"},{"alias_kind":"pith_short_8","alias_value":"KNJVBENM","created_at":"2026-07-05T10:06:02.965989+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2404.02138","citing_title":"Topic-Based Watermarks for Large Language Models","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12456","citing_title":"TextSeal: A Localized LLM Watermark for Provenance & Distillation Protection","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12456","citing_title":"TextSeal: A Localized LLM Watermark for Provenance & Distillation Protection","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11653","citing_title":"Every Bit, Everywhere, All at Once: A Binomial Multibit LLM Watermark","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08964","citing_title":"Trustworthy AI: Ensuring Reliability and Accountability from Models to Agents","ref_index":133,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KNJVBENMTS3USTJJWC6GX5KK5F","json":"https://pith.science/pith/KNJVBENMTS3USTJJWC6GX5KK5F.json","graph_json":"https://pith.science/api/pith-number/KNJVBENMTS3USTJJWC6GX5KK5F/graph.json","events_json":"https://pith.science/api/pith-number/KNJVBENMTS3USTJJWC6GX5KK5F/events.json","paper":"https://pith.science/paper/KNJVBENM"},"agent_actions":{"view_html":"https://pith.science/pith/KNJVBENMTS3USTJJWC6GX5KK5F","download_json":"https://pith.science/pith/KNJVBENMTS3USTJJWC6GX5KK5F.json","view_paper":"https://pith.science/paper/KNJVBENM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.16820&json=true","fetch_graph":"https://pith.science/api/pith-number/KNJVBENMTS3USTJJWC6GX5KK5F/graph.json","fetch_events":"https://pith.science/api/pith-number/KNJVBENMTS3USTJJWC6GX5KK5F/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KNJVBENMTS3USTJJWC6GX5KK5F/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KNJVBENMTS3USTJJWC6GX5KK5F/action/storage_attestation","attest_author":"https://pith.science/pith/KNJVBENMTS3USTJJWC6GX5KK5F/action/author_attestation","sign_citation":"https://pith.science/pith/KNJVBENMTS3USTJJWC6GX5KK5F/action/citation_signature","submit_replication":"https://pith.science/pith/KNJVBENMTS3USTJJWC6GX5KK5F/action/replication_record"}},"created_at":"2026-07-05T10:06:02.965989+00:00","updated_at":"2026-07-05T10:06:02.965989+00:00"}