{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ZQ64VWY36MA5P3YL7QFCLJ5RWL","short_pith_number":"pith:ZQ64VWY3","schema_version":"1.0","canonical_sha256":"cc3dcadb1bf301d7ef0bfc0a25a7b1b2fd280775cfc6975fbc089e4739f8b9c0","source":{"kind":"arxiv","id":"2509.03058","version":1},"attestation_state":"computed","paper":{"title":"EverTracer: Hunting Stolen Large Language Models via Stealthy and Robust Probabilistic Fingerprint","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Meng Han, Wenpeng Xing, Zhenhua Xu","submitted_at":"2025-09-03T06:40:57Z","abstract_excerpt":"The proliferation of large language models (LLMs) has intensified concerns over model theft and license violations, necessitating robust and stealthy ownership verification. Existing fingerprinting methods either require impractical white-box access or introduce detectable statistical anomalies. We propose EverTracer, a novel gray-box fingerprinting framework that ensures stealthy and robust model provenance tracing. EverTracer is the first to repurpose Membership Inference Attacks (MIAs) for defensive use, embedding ownership signals via memorization instead of artificial trigger-output overf"},"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":"2509.03058","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2025-09-03T06:40:57Z","cross_cats_sorted":[],"title_canon_sha256":"ab4d57f7e4d4b9f04caa462662dfe1dd64014196c4205ed5e51c7706425722b3","abstract_canon_sha256":"188282e44d73a329186c00a0307a73e31b435c4a6a86dd3809ec577be89024d7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:04:02.996109Z","signature_b64":"moOJLP/SqC1FXz6eCKrgkNpScql1G7Wj7ceXSjnxBzoFGqFa43MNUqmvk+IZR3AYUY1OgMdKNm93xbvCxlBtBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cc3dcadb1bf301d7ef0bfc0a25a7b1b2fd280775cfc6975fbc089e4739f8b9c0","last_reissued_at":"2026-07-05T12:04:02.995626Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:04:02.995626Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"EverTracer: Hunting Stolen Large Language Models via Stealthy and Robust Probabilistic Fingerprint","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Meng Han, Wenpeng Xing, Zhenhua Xu","submitted_at":"2025-09-03T06:40:57Z","abstract_excerpt":"The proliferation of large language models (LLMs) has intensified concerns over model theft and license violations, necessitating robust and stealthy ownership verification. Existing fingerprinting methods either require impractical white-box access or introduce detectable statistical anomalies. We propose EverTracer, a novel gray-box fingerprinting framework that ensures stealthy and robust model provenance tracing. EverTracer is the first to repurpose Membership Inference Attacks (MIAs) for defensive use, embedding ownership signals via memorization instead of artificial trigger-output overf"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.03058","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/2509.03058/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":"2509.03058","created_at":"2026-07-05T12:04:02.995683+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.03058v1","created_at":"2026-07-05T12:04:02.995683+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.03058","created_at":"2026-07-05T12:04:02.995683+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZQ64VWY36MA5","created_at":"2026-07-05T12:04:02.995683+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZQ64VWY36MA5P3YL","created_at":"2026-07-05T12:04:02.995683+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZQ64VWY3","created_at":"2026-07-05T12:04:02.995683+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2509.09703","citing_title":"CTCC: A Robust and Stealthy Fingerprinting Framework for Large Language Models via Cross-Turn Contextual Correlation Backdoor","ref_index":45,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZQ64VWY36MA5P3YL7QFCLJ5RWL","json":"https://pith.science/pith/ZQ64VWY36MA5P3YL7QFCLJ5RWL.json","graph_json":"https://pith.science/api/pith-number/ZQ64VWY36MA5P3YL7QFCLJ5RWL/graph.json","events_json":"https://pith.science/api/pith-number/ZQ64VWY36MA5P3YL7QFCLJ5RWL/events.json","paper":"https://pith.science/paper/ZQ64VWY3"},"agent_actions":{"view_html":"https://pith.science/pith/ZQ64VWY36MA5P3YL7QFCLJ5RWL","download_json":"https://pith.science/pith/ZQ64VWY36MA5P3YL7QFCLJ5RWL.json","view_paper":"https://pith.science/paper/ZQ64VWY3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.03058&json=true","fetch_graph":"https://pith.science/api/pith-number/ZQ64VWY36MA5P3YL7QFCLJ5RWL/graph.json","fetch_events":"https://pith.science/api/pith-number/ZQ64VWY36MA5P3YL7QFCLJ5RWL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZQ64VWY36MA5P3YL7QFCLJ5RWL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZQ64VWY36MA5P3YL7QFCLJ5RWL/action/storage_attestation","attest_author":"https://pith.science/pith/ZQ64VWY36MA5P3YL7QFCLJ5RWL/action/author_attestation","sign_citation":"https://pith.science/pith/ZQ64VWY36MA5P3YL7QFCLJ5RWL/action/citation_signature","submit_replication":"https://pith.science/pith/ZQ64VWY36MA5P3YL7QFCLJ5RWL/action/replication_record"}},"created_at":"2026-07-05T12:04:02.995683+00:00","updated_at":"2026-07-05T12:04:02.995683+00:00"}