{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:X4RLKOE5XVPERC6XITDBLIAOVQ","short_pith_number":"pith:X4RLKOE5","schema_version":"1.0","canonical_sha256":"bf22b5389dbd5e488bd744c615a00eac2d2d5f99d62acc19b73ff3447db9e18e","source":{"kind":"arxiv","id":"2505.12682","version":1},"attestation_state":"computed","paper":{"title":"RoFL: Robust Fingerprinting of Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chuan Guo, Junfeng Yang, Laurens van der Maaten, Yun-Yun Tsai","submitted_at":"2025-05-19T04:00:23Z","abstract_excerpt":"AI developers are releasing large language models (LLMs) under a variety of different licenses. Many of these licenses restrict the ways in which the models or their outputs may be used. This raises the question how license violations may be recognized. In particular, how can we identify that an API or product uses (an adapted version of) a particular LLM? We present a new method that enable model developers to perform such identification via fingerprints: statistical patterns that are unique to the developer's model and robust to common alterations of that model. Our method permits model iden"},"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":"2505.12682","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-19T04:00:23Z","cross_cats_sorted":[],"title_canon_sha256":"977067997cfe7037dce66adafc85544f6cc38049d157490f534424a4e74fc778","abstract_canon_sha256":"b96f006c1e8c848838987e4db08751174a9c758555f86046d211f236fae8a69f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:05:08.126582Z","signature_b64":"zCWCvH8J4OvDyMjuLMCt4U7Yl/mDdy7GtcZan2HWWbM2jLaEkRcYGRVekLDLAS1nGYkvPEayoKnHREL/3W61Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bf22b5389dbd5e488bd744c615a00eac2d2d5f99d62acc19b73ff3447db9e18e","last_reissued_at":"2026-07-05T11:05:08.126085Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:05:08.126085Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"RoFL: Robust Fingerprinting of Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chuan Guo, Junfeng Yang, Laurens van der Maaten, Yun-Yun Tsai","submitted_at":"2025-05-19T04:00:23Z","abstract_excerpt":"AI developers are releasing large language models (LLMs) under a variety of different licenses. Many of these licenses restrict the ways in which the models or their outputs may be used. This raises the question how license violations may be recognized. In particular, how can we identify that an API or product uses (an adapted version of) a particular LLM? We present a new method that enable model developers to perform such identification via fingerprints: statistical patterns that are unique to the developer's model and robust to common alterations of that model. Our method permits model iden"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.12682","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/2505.12682/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":"2505.12682","created_at":"2026-07-05T11:05:08.126142+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.12682v1","created_at":"2026-07-05T11:05:08.126142+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.12682","created_at":"2026-07-05T11:05:08.126142+00:00"},{"alias_kind":"pith_short_12","alias_value":"X4RLKOE5XVPE","created_at":"2026-07-05T11:05:08.126142+00:00"},{"alias_kind":"pith_short_16","alias_value":"X4RLKOE5XVPERC6X","created_at":"2026-07-05T11:05:08.126142+00:00"},{"alias_kind":"pith_short_8","alias_value":"X4RLKOE5","created_at":"2026-07-05T11:05:08.126142+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":5,"sample":[{"citing_arxiv_id":"2606.03330","citing_title":"FLIPS: Instance-Fingerprinting for LLMs via Pseudo-random Sequences","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2605.29524","citing_title":"KBF: Knowledge Boundary as Fingerprint for Language Model and Black-Box API Auditing","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2508.11548","citing_title":"Copyright Protection for Large Language Models: A Survey of Methods, Challenges, and Trends","ref_index":142,"is_internal_anchor":true},{"citing_arxiv_id":"2509.24496","citing_title":"LLM DNA: Tracing Model Evolution via Functional Representations","ref_index":12,"is_internal_anchor":true},{"citing_arxiv_id":"2509.26404","citing_title":"SeedPrints: Fingerprints Can Even Tell Which Seed Your Large Language Model Was Trained From","ref_index":10,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/X4RLKOE5XVPERC6XITDBLIAOVQ","json":"https://pith.science/pith/X4RLKOE5XVPERC6XITDBLIAOVQ.json","graph_json":"https://pith.science/api/pith-number/X4RLKOE5XVPERC6XITDBLIAOVQ/graph.json","events_json":"https://pith.science/api/pith-number/X4RLKOE5XVPERC6XITDBLIAOVQ/events.json","paper":"https://pith.science/paper/X4RLKOE5"},"agent_actions":{"view_html":"https://pith.science/pith/X4RLKOE5XVPERC6XITDBLIAOVQ","download_json":"https://pith.science/pith/X4RLKOE5XVPERC6XITDBLIAOVQ.json","view_paper":"https://pith.science/paper/X4RLKOE5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.12682&json=true","fetch_graph":"https://pith.science/api/pith-number/X4RLKOE5XVPERC6XITDBLIAOVQ/graph.json","fetch_events":"https://pith.science/api/pith-number/X4RLKOE5XVPERC6XITDBLIAOVQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/X4RLKOE5XVPERC6XITDBLIAOVQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/X4RLKOE5XVPERC6XITDBLIAOVQ/action/storage_attestation","attest_author":"https://pith.science/pith/X4RLKOE5XVPERC6XITDBLIAOVQ/action/author_attestation","sign_citation":"https://pith.science/pith/X4RLKOE5XVPERC6XITDBLIAOVQ/action/citation_signature","submit_replication":"https://pith.science/pith/X4RLKOE5XVPERC6XITDBLIAOVQ/action/replication_record"}},"created_at":"2026-07-05T11:05:08.126142+00:00","updated_at":"2026-07-05T11:05:08.126142+00:00"}