{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:BZ7ASGOKDJXR2H6C5Q64A2QGMA","short_pith_number":"pith:BZ7ASGOK","schema_version":"1.0","canonical_sha256":"0e7e0919ca1a6f1d1fc2ec3dc06a0660194fddb56de6fa8bd9f3b9caf3994e1c","source":{"kind":"arxiv","id":"2403.00871","version":1},"attestation_state":"computed","paper":{"title":"Teach LLMs to Phish: Stealing Private Information from Language Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.CR","authors_text":"Ashwinee Panda, Christopher A. Choquette-Choo, Prateek Mittal, Yaoqing Yang, Zhengming Zhang","submitted_at":"2024-03-01T06:15:07Z","abstract_excerpt":"When large language models are trained on private data, it can be a significant privacy risk for them to memorize and regurgitate sensitive information. In this work, we propose a new practical data extraction attack that we call \"neural phishing\". This attack enables an adversary to target and extract sensitive or personally identifiable information (PII), e.g., credit card numbers, from a model trained on user data with upwards of 10% attack success rates, at times, as high as 50%. Our attack assumes only that an adversary can insert as few as 10s of benign-appearing sentences into the train"},"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":"2403.00871","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2024-03-01T06:15:07Z","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"title_canon_sha256":"8337aa71261007f697dd0b4442e7299b6ece0f3d05f128dec93b8716322cac21","abstract_canon_sha256":"64e15c7a807f2cc352998e449cdd8a3c4f3fda70fbcdae9dcf1ffc9bec7ac44b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:51:13.048156Z","signature_b64":"Kj5COYD+KhvOIU2/ronw9zmSH1PXf1BxwYh5C3udhr4TtQluID0GHd64aiZ97a4fhpLb5WnO+jObTNK/sCArAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0e7e0919ca1a6f1d1fc2ec3dc06a0660194fddb56de6fa8bd9f3b9caf3994e1c","last_reissued_at":"2026-07-05T07:51:13.047733Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:51:13.047733Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Teach LLMs to Phish: Stealing Private Information from Language Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.CR","authors_text":"Ashwinee Panda, Christopher A. Choquette-Choo, Prateek Mittal, Yaoqing Yang, Zhengming Zhang","submitted_at":"2024-03-01T06:15:07Z","abstract_excerpt":"When large language models are trained on private data, it can be a significant privacy risk for them to memorize and regurgitate sensitive information. In this work, we propose a new practical data extraction attack that we call \"neural phishing\". This attack enables an adversary to target and extract sensitive or personally identifiable information (PII), e.g., credit card numbers, from a model trained on user data with upwards of 10% attack success rates, at times, as high as 50%. Our attack assumes only that an adversary can insert as few as 10s of benign-appearing sentences into the train"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.00871","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/2403.00871/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":"2403.00871","created_at":"2026-07-05T07:51:13.047792+00:00"},{"alias_kind":"arxiv_version","alias_value":"2403.00871v1","created_at":"2026-07-05T07:51:13.047792+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.00871","created_at":"2026-07-05T07:51:13.047792+00:00"},{"alias_kind":"pith_short_12","alias_value":"BZ7ASGOKDJXR","created_at":"2026-07-05T07:51:13.047792+00:00"},{"alias_kind":"pith_short_16","alias_value":"BZ7ASGOKDJXR2H6C","created_at":"2026-07-05T07:51:13.047792+00:00"},{"alias_kind":"pith_short_8","alias_value":"BZ7ASGOK","created_at":"2026-07-05T07:51:13.047792+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2508.21457","citing_title":"SoK: Exposing the Generation and Detection Gaps in LLM-Generated Phishing","ref_index":71,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BZ7ASGOKDJXR2H6C5Q64A2QGMA","json":"https://pith.science/pith/BZ7ASGOKDJXR2H6C5Q64A2QGMA.json","graph_json":"https://pith.science/api/pith-number/BZ7ASGOKDJXR2H6C5Q64A2QGMA/graph.json","events_json":"https://pith.science/api/pith-number/BZ7ASGOKDJXR2H6C5Q64A2QGMA/events.json","paper":"https://pith.science/paper/BZ7ASGOK"},"agent_actions":{"view_html":"https://pith.science/pith/BZ7ASGOKDJXR2H6C5Q64A2QGMA","download_json":"https://pith.science/pith/BZ7ASGOKDJXR2H6C5Q64A2QGMA.json","view_paper":"https://pith.science/paper/BZ7ASGOK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2403.00871&json=true","fetch_graph":"https://pith.science/api/pith-number/BZ7ASGOKDJXR2H6C5Q64A2QGMA/graph.json","fetch_events":"https://pith.science/api/pith-number/BZ7ASGOKDJXR2H6C5Q64A2QGMA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BZ7ASGOKDJXR2H6C5Q64A2QGMA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BZ7ASGOKDJXR2H6C5Q64A2QGMA/action/storage_attestation","attest_author":"https://pith.science/pith/BZ7ASGOKDJXR2H6C5Q64A2QGMA/action/author_attestation","sign_citation":"https://pith.science/pith/BZ7ASGOKDJXR2H6C5Q64A2QGMA/action/citation_signature","submit_replication":"https://pith.science/pith/BZ7ASGOKDJXR2H6C5Q64A2QGMA/action/replication_record"}},"created_at":"2026-07-05T07:51:13.047792+00:00","updated_at":"2026-07-05T07:51:13.047792+00:00"}