{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:KUVHJGEQ5ITKGR5UGOGRYNPUZF","short_pith_number":"pith:KUVHJGEQ","schema_version":"1.0","canonical_sha256":"552a749890ea26a347b4338d1c35f4c952b52fe1f98722a767e18546c2eb8b4b","source":{"kind":"arxiv","id":"2405.19677","version":1},"attestation_state":"computed","paper":{"title":"Large Language Model Watermark Stealing With Mixed Integer Programming","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Asif Gill, Chao Chen, Leo Yu Zhang, Shengshan Hu, Shirui Pan, Xiaomei Zhang, Yanjun Zhang, Zhaoxi Zhang","submitted_at":"2024-05-30T04:11:17Z","abstract_excerpt":"The Large Language Model (LLM) watermark is a newly emerging technique that shows promise in addressing concerns surrounding LLM copyright, monitoring AI-generated text, and preventing its misuse. The LLM watermark scheme commonly includes generating secret keys to partition the vocabulary into green and red lists, applying a perturbation to the logits of tokens in the green list to increase their sampling likelihood, thus facilitating watermark detection to identify AI-generated text if the proportion of green tokens exceeds a threshold. However, recent research indicates that watermarking me"},"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":"2405.19677","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CR","submitted_at":"2024-05-30T04:11:17Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8d61b5875b98f9f85a6054a1457698cbc239e08151280936b152936f5e47a51e","abstract_canon_sha256":"0e8463ab71a6a4488c096b2a631dfee943d942e0fdd628616f9308001688ce21"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:25:11.064411Z","signature_b64":"eZj1qFt0gUrxSiCI5dlhP4JAMDlmfVYV53L5jhY118mZoFY2iWD73qjFQ93TwPwqtqEHbIYugOi++Gy/0ruWDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"552a749890ea26a347b4338d1c35f4c952b52fe1f98722a767e18546c2eb8b4b","last_reissued_at":"2026-07-05T08:25:11.063952Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:25:11.063952Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Large Language Model Watermark Stealing With Mixed Integer Programming","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Asif Gill, Chao Chen, Leo Yu Zhang, Shengshan Hu, Shirui Pan, Xiaomei Zhang, Yanjun Zhang, Zhaoxi Zhang","submitted_at":"2024-05-30T04:11:17Z","abstract_excerpt":"The Large Language Model (LLM) watermark is a newly emerging technique that shows promise in addressing concerns surrounding LLM copyright, monitoring AI-generated text, and preventing its misuse. The LLM watermark scheme commonly includes generating secret keys to partition the vocabulary into green and red lists, applying a perturbation to the logits of tokens in the green list to increase their sampling likelihood, thus facilitating watermark detection to identify AI-generated text if the proportion of green tokens exceeds a threshold. However, recent research indicates that watermarking me"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.19677","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/2405.19677/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":"2405.19677","created_at":"2026-07-05T08:25:11.064009+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.19677v1","created_at":"2026-07-05T08:25:11.064009+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.19677","created_at":"2026-07-05T08:25:11.064009+00:00"},{"alias_kind":"pith_short_12","alias_value":"KUVHJGEQ5ITK","created_at":"2026-07-05T08:25:11.064009+00:00"},{"alias_kind":"pith_short_16","alias_value":"KUVHJGEQ5ITKGR5U","created_at":"2026-07-05T08:25:11.064009+00:00"},{"alias_kind":"pith_short_8","alias_value":"KUVHJGEQ","created_at":"2026-07-05T08:25:11.064009+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"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":29,"is_internal_anchor":false},{"citing_arxiv_id":"2507.07871","citing_title":"Mitigating Watermark Forgery in Generative Models via Randomized Key Selection","ref_index":46,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13095","citing_title":"Watermarking Should Be Treated as a Monitoring Primitive","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13095","citing_title":"Watermarking Should Be Treated as a Monitoring Primitive","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2604.06662","citing_title":"Towards Robust Content Watermarking Against Removal and Forgery Attacks","ref_index":68,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KUVHJGEQ5ITKGR5UGOGRYNPUZF","json":"https://pith.science/pith/KUVHJGEQ5ITKGR5UGOGRYNPUZF.json","graph_json":"https://pith.science/api/pith-number/KUVHJGEQ5ITKGR5UGOGRYNPUZF/graph.json","events_json":"https://pith.science/api/pith-number/KUVHJGEQ5ITKGR5UGOGRYNPUZF/events.json","paper":"https://pith.science/paper/KUVHJGEQ"},"agent_actions":{"view_html":"https://pith.science/pith/KUVHJGEQ5ITKGR5UGOGRYNPUZF","download_json":"https://pith.science/pith/KUVHJGEQ5ITKGR5UGOGRYNPUZF.json","view_paper":"https://pith.science/paper/KUVHJGEQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.19677&json=true","fetch_graph":"https://pith.science/api/pith-number/KUVHJGEQ5ITKGR5UGOGRYNPUZF/graph.json","fetch_events":"https://pith.science/api/pith-number/KUVHJGEQ5ITKGR5UGOGRYNPUZF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KUVHJGEQ5ITKGR5UGOGRYNPUZF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KUVHJGEQ5ITKGR5UGOGRYNPUZF/action/storage_attestation","attest_author":"https://pith.science/pith/KUVHJGEQ5ITKGR5UGOGRYNPUZF/action/author_attestation","sign_citation":"https://pith.science/pith/KUVHJGEQ5ITKGR5UGOGRYNPUZF/action/citation_signature","submit_replication":"https://pith.science/pith/KUVHJGEQ5ITKGR5UGOGRYNPUZF/action/replication_record"}},"created_at":"2026-07-05T08:25:11.064009+00:00","updated_at":"2026-07-05T08:25:11.064009+00:00"}