{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:WVWK52DFBNVSNUCWX5TDYAU7JB","short_pith_number":"pith:WVWK52DF","schema_version":"1.0","canonical_sha256":"b56caee8650b6b26d056bf663c029f486f5db31ee04fdbc2b6bcc9af3d375952","source":{"kind":"arxiv","id":"2305.08883","version":1},"attestation_state":"computed","paper":{"title":"Watermarking Text Generated by Black-Box Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chang Liu, Han Fang, Jie Zhang, Kejiang Chen, Nenghai Yu, Weiming Zhang, Xi Yang, Yuang Qi","submitted_at":"2023-05-14T07:37:33Z","abstract_excerpt":"LLMs now exhibit human-like skills in various fields, leading to worries about misuse. Thus, detecting generated text is crucial. However, passive detection methods are stuck in domain specificity and limited adversarial robustness. To achieve reliable detection, a watermark-based method was proposed for white-box LLMs, allowing them to embed watermarks during text generation. The method involves randomly dividing the model vocabulary to obtain a special list and adjusting the probability distribution to promote the selection of words in the list. A detection algorithm aware of the list can id"},"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":"2305.08883","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-14T07:37:33Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"78d0554030c937752f693dd17c14e6a06f9ed5906c97417bc6df0d0aadd960af","abstract_canon_sha256":"76c591e4bdafa3a0b73d9175924912af7a4b1b904f3ba4f930c2589e30af9dec"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:10:41.410575Z","signature_b64":"ZXDTv3eTI51TQ2hulIOvNrIlLaW86lJT5gKp1+jjivtIgyoVSuQIByn/F/MCXiNJFuay6wv388b2zyBA8p2nCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b56caee8650b6b26d056bf663c029f486f5db31ee04fdbc2b6bcc9af3d375952","last_reissued_at":"2026-07-05T06:10:41.410155Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:10:41.410155Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Watermarking Text Generated by Black-Box Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chang Liu, Han Fang, Jie Zhang, Kejiang Chen, Nenghai Yu, Weiming Zhang, Xi Yang, Yuang Qi","submitted_at":"2023-05-14T07:37:33Z","abstract_excerpt":"LLMs now exhibit human-like skills in various fields, leading to worries about misuse. Thus, detecting generated text is crucial. However, passive detection methods are stuck in domain specificity and limited adversarial robustness. To achieve reliable detection, a watermark-based method was proposed for white-box LLMs, allowing them to embed watermarks during text generation. The method involves randomly dividing the model vocabulary to obtain a special list and adjusting the probability distribution to promote the selection of words in the list. A detection algorithm aware of the list can id"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.08883","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/2305.08883/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":"2305.08883","created_at":"2026-07-05T06:10:41.410209+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.08883v1","created_at":"2026-07-05T06:10:41.410209+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.08883","created_at":"2026-07-05T06:10:41.410209+00:00"},{"alias_kind":"pith_short_12","alias_value":"WVWK52DFBNVS","created_at":"2026-07-05T06:10:41.410209+00:00"},{"alias_kind":"pith_short_16","alias_value":"WVWK52DFBNVSNUCW","created_at":"2026-07-05T06:10:41.410209+00:00"},{"alias_kind":"pith_short_8","alias_value":"WVWK52DF","created_at":"2026-07-05T06:10:41.410209+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.05353","citing_title":"Selective Disclosure Watermarking for Large Language Models","ref_index":11,"is_internal_anchor":true},{"citing_arxiv_id":"2606.00613","citing_title":"Linguistics-Aware Non-Distortionary LLM Watermarking","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2508.11548","citing_title":"Copyright Protection for Large Language Models: A Survey of Methods, Challenges, and Trends","ref_index":169,"is_internal_anchor":false},{"citing_arxiv_id":"2510.18333","citing_title":"Position: LLM Watermarking Should Align Stakeholders' Incentives for Practical Adoption","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08964","citing_title":"Trustworthy AI: Ensuring Reliability and Accountability from Models to Agents","ref_index":164,"is_internal_anchor":false},{"citing_arxiv_id":"2604.25247","citing_title":"R-CoT: A Reasoning-Layer Watermark via Redundant Chain-of-Thought in Large Language Models","ref_index":27,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WVWK52DFBNVSNUCWX5TDYAU7JB","json":"https://pith.science/pith/WVWK52DFBNVSNUCWX5TDYAU7JB.json","graph_json":"https://pith.science/api/pith-number/WVWK52DFBNVSNUCWX5TDYAU7JB/graph.json","events_json":"https://pith.science/api/pith-number/WVWK52DFBNVSNUCWX5TDYAU7JB/events.json","paper":"https://pith.science/paper/WVWK52DF"},"agent_actions":{"view_html":"https://pith.science/pith/WVWK52DFBNVSNUCWX5TDYAU7JB","download_json":"https://pith.science/pith/WVWK52DFBNVSNUCWX5TDYAU7JB.json","view_paper":"https://pith.science/paper/WVWK52DF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.08883&json=true","fetch_graph":"https://pith.science/api/pith-number/WVWK52DFBNVSNUCWX5TDYAU7JB/graph.json","fetch_events":"https://pith.science/api/pith-number/WVWK52DFBNVSNUCWX5TDYAU7JB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WVWK52DFBNVSNUCWX5TDYAU7JB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WVWK52DFBNVSNUCWX5TDYAU7JB/action/storage_attestation","attest_author":"https://pith.science/pith/WVWK52DFBNVSNUCWX5TDYAU7JB/action/author_attestation","sign_citation":"https://pith.science/pith/WVWK52DFBNVSNUCWX5TDYAU7JB/action/citation_signature","submit_replication":"https://pith.science/pith/WVWK52DFBNVSNUCWX5TDYAU7JB/action/replication_record"}},"created_at":"2026-07-05T06:10:41.410209+00:00","updated_at":"2026-07-05T06:10:41.410209+00:00"}