{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:BSZVHPZE6NIMLN3L2F3WLNMUPQ","short_pith_number":"pith:BSZVHPZE","schema_version":"1.0","canonical_sha256":"0cb353bf24f350c5b76bd17765b5947c30c88c55e3d70ea937e33bdafaad6ec0","source":{"kind":"arxiv","id":"2307.04401","version":1},"attestation_state":"computed","paper":{"title":"Ethicist: Targeted Training Data Extraction Through Loss Smoothed Soft Prompting and Calibrated Confidence Estimation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Jiaxin Wen, Minlie Huang, Zhexin Zhang","submitted_at":"2023-07-10T08:03:41Z","abstract_excerpt":"Large pre-trained language models achieve impressive results across many tasks. However, recent works point out that pre-trained language models may memorize a considerable fraction of their training data, leading to the privacy risk of information leakage. In this paper, we propose a method named Ethicist for targeted training data extraction through loss smoothed soft prompting and calibrated confidence estimation, investigating how to recover the suffix in the training data when given a prefix. To elicit memorization in the attacked model, we tune soft prompt embeddings while keeping the mo"},"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":"2307.04401","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-07-10T08:03:41Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ab067326313299037b653930c5de816e3c4ad156d07bf4784b69e5e27a45520b","abstract_canon_sha256":"27fc24651a76c7b39e247228abb213c2dc363fd0c71204d6ccd0881d0cce7eef"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:29:12.690377Z","signature_b64":"QUe63c3MAfEZZ8iZV9uPU1T4q+YsL9lC5h0mQAV5JtH75/loED6+2HlbFnouxCyl0wrtwgVeyoo1PZ1D8+gADQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0cb353bf24f350c5b76bd17765b5947c30c88c55e3d70ea937e33bdafaad6ec0","last_reissued_at":"2026-07-05T06:29:12.689905Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:29:12.689905Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Ethicist: Targeted Training Data Extraction Through Loss Smoothed Soft Prompting and Calibrated Confidence Estimation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Jiaxin Wen, Minlie Huang, Zhexin Zhang","submitted_at":"2023-07-10T08:03:41Z","abstract_excerpt":"Large pre-trained language models achieve impressive results across many tasks. However, recent works point out that pre-trained language models may memorize a considerable fraction of their training data, leading to the privacy risk of information leakage. In this paper, we propose a method named Ethicist for targeted training data extraction through loss smoothed soft prompting and calibrated confidence estimation, investigating how to recover the suffix in the training data when given a prefix. To elicit memorization in the attacked model, we tune soft prompt embeddings while keeping the mo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2307.04401","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/2307.04401/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":"2307.04401","created_at":"2026-07-05T06:29:12.689964+00:00"},{"alias_kind":"arxiv_version","alias_value":"2307.04401v1","created_at":"2026-07-05T06:29:12.689964+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2307.04401","created_at":"2026-07-05T06:29:12.689964+00:00"},{"alias_kind":"pith_short_12","alias_value":"BSZVHPZE6NIM","created_at":"2026-07-05T06:29:12.689964+00:00"},{"alias_kind":"pith_short_16","alias_value":"BSZVHPZE6NIMLN3L","created_at":"2026-07-05T06:29:12.689964+00:00"},{"alias_kind":"pith_short_8","alias_value":"BSZVHPZE","created_at":"2026-07-05T06:29:12.689964+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.15031","citing_title":"A Systematic Survey of Model Extraction Attacks and Defenses: State-of-the-Art and Perspectives","ref_index":261,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BSZVHPZE6NIMLN3L2F3WLNMUPQ","json":"https://pith.science/pith/BSZVHPZE6NIMLN3L2F3WLNMUPQ.json","graph_json":"https://pith.science/api/pith-number/BSZVHPZE6NIMLN3L2F3WLNMUPQ/graph.json","events_json":"https://pith.science/api/pith-number/BSZVHPZE6NIMLN3L2F3WLNMUPQ/events.json","paper":"https://pith.science/paper/BSZVHPZE"},"agent_actions":{"view_html":"https://pith.science/pith/BSZVHPZE6NIMLN3L2F3WLNMUPQ","download_json":"https://pith.science/pith/BSZVHPZE6NIMLN3L2F3WLNMUPQ.json","view_paper":"https://pith.science/paper/BSZVHPZE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2307.04401&json=true","fetch_graph":"https://pith.science/api/pith-number/BSZVHPZE6NIMLN3L2F3WLNMUPQ/graph.json","fetch_events":"https://pith.science/api/pith-number/BSZVHPZE6NIMLN3L2F3WLNMUPQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BSZVHPZE6NIMLN3L2F3WLNMUPQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BSZVHPZE6NIMLN3L2F3WLNMUPQ/action/storage_attestation","attest_author":"https://pith.science/pith/BSZVHPZE6NIMLN3L2F3WLNMUPQ/action/author_attestation","sign_citation":"https://pith.science/pith/BSZVHPZE6NIMLN3L2F3WLNMUPQ/action/citation_signature","submit_replication":"https://pith.science/pith/BSZVHPZE6NIMLN3L2F3WLNMUPQ/action/replication_record"}},"created_at":"2026-07-05T06:29:12.689964+00:00","updated_at":"2026-07-05T06:29:12.689964+00:00"}