{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:YO4MA6G4BPWDOLDX35TS7T2GCP","short_pith_number":"pith:YO4MA6G4","schema_version":"1.0","canonical_sha256":"c3b8c078dc0bec372c77df672fcf4613d18b11be249206fb7b32c872c07bb921","source":{"kind":"arxiv","id":"2402.14007","version":2},"attestation_state":"computed","paper":{"title":"Can Watermarks Survive Translation? On the Cross-lingual Consistency of Text Watermark for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aiwei Liu, Binglin Zhou, Hongkun Hao, Rui Wang, Xing Wang, Zhaopeng Tu, Zhiwei He, Zhuosheng Zhang","submitted_at":"2024-02-21T18:48:38Z","abstract_excerpt":"Text watermarking technology aims to tag and identify content produced by large language models (LLMs) to prevent misuse. In this study, we introduce the concept of cross-lingual consistency in text watermarking, which assesses the ability of text watermarks to maintain their effectiveness after being translated into other languages. Preliminary empirical results from two LLMs and three watermarking methods reveal that current text watermarking technologies lack consistency when texts are translated into various languages. Based on this observation, we propose a Cross-lingual Watermark Removal"},"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":"2402.14007","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-02-21T18:48:38Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ddfdd6cd9f245f7b640952e9105b5352a6ac409cc3c52d7c4b44f310f576e033","abstract_canon_sha256":"2ab7336f508b02dd60673b994d7ee2e9ac4a0a35c97534b34e7a59109d6b942d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:27:04.476890Z","signature_b64":"pMrwambscBsBcQrp0pn5YBEUr4zeiMJlV8huo5/rLZgkCCmAzHGhmNmYZOX0oN/Dv2BYysaOdAoNptC0urBAAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c3b8c078dc0bec372c77df672fcf4613d18b11be249206fb7b32c872c07bb921","last_reissued_at":"2026-07-05T08:27:04.476396Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:27:04.476396Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Can Watermarks Survive Translation? On the Cross-lingual Consistency of Text Watermark for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aiwei Liu, Binglin Zhou, Hongkun Hao, Rui Wang, Xing Wang, Zhaopeng Tu, Zhiwei He, Zhuosheng Zhang","submitted_at":"2024-02-21T18:48:38Z","abstract_excerpt":"Text watermarking technology aims to tag and identify content produced by large language models (LLMs) to prevent misuse. In this study, we introduce the concept of cross-lingual consistency in text watermarking, which assesses the ability of text watermarks to maintain their effectiveness after being translated into other languages. Preliminary empirical results from two LLMs and three watermarking methods reveal that current text watermarking technologies lack consistency when texts are translated into various languages. Based on this observation, we propose a Cross-lingual Watermark Removal"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.14007","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/2402.14007/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":"2402.14007","created_at":"2026-07-05T08:27:04.476456+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.14007v2","created_at":"2026-07-05T08:27:04.476456+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.14007","created_at":"2026-07-05T08:27:04.476456+00:00"},{"alias_kind":"pith_short_12","alias_value":"YO4MA6G4BPWD","created_at":"2026-07-05T08:27:04.476456+00:00"},{"alias_kind":"pith_short_16","alias_value":"YO4MA6G4BPWDOLDX","created_at":"2026-07-05T08:27:04.476456+00:00"},{"alias_kind":"pith_short_8","alias_value":"YO4MA6G4","created_at":"2026-07-05T08:27:04.476456+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18430","citing_title":"Signature filtering: a lightweight enhancement for statistical watermark detection in large language models","ref_index":14,"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":56,"is_internal_anchor":false},{"citing_arxiv_id":"2509.20924","citing_title":"RLCracker: Evaluating the Worst-Case Vulnerability of LLM Watermarks with Adaptive RL Attacks","ref_index":7,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YO4MA6G4BPWDOLDX35TS7T2GCP","json":"https://pith.science/pith/YO4MA6G4BPWDOLDX35TS7T2GCP.json","graph_json":"https://pith.science/api/pith-number/YO4MA6G4BPWDOLDX35TS7T2GCP/graph.json","events_json":"https://pith.science/api/pith-number/YO4MA6G4BPWDOLDX35TS7T2GCP/events.json","paper":"https://pith.science/paper/YO4MA6G4"},"agent_actions":{"view_html":"https://pith.science/pith/YO4MA6G4BPWDOLDX35TS7T2GCP","download_json":"https://pith.science/pith/YO4MA6G4BPWDOLDX35TS7T2GCP.json","view_paper":"https://pith.science/paper/YO4MA6G4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.14007&json=true","fetch_graph":"https://pith.science/api/pith-number/YO4MA6G4BPWDOLDX35TS7T2GCP/graph.json","fetch_events":"https://pith.science/api/pith-number/YO4MA6G4BPWDOLDX35TS7T2GCP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YO4MA6G4BPWDOLDX35TS7T2GCP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YO4MA6G4BPWDOLDX35TS7T2GCP/action/storage_attestation","attest_author":"https://pith.science/pith/YO4MA6G4BPWDOLDX35TS7T2GCP/action/author_attestation","sign_citation":"https://pith.science/pith/YO4MA6G4BPWDOLDX35TS7T2GCP/action/citation_signature","submit_replication":"https://pith.science/pith/YO4MA6G4BPWDOLDX35TS7T2GCP/action/replication_record"}},"created_at":"2026-07-05T08:27:04.476456+00:00","updated_at":"2026-07-05T08:27:04.476456+00:00"}