{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:467VBC5WZ23L4ZLS7PGXMR2XMU","short_pith_number":"pith:467VBC5W","canonical_record":{"source":{"id":"2505.06284","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-07T07:21:37Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"6297b2def7b58dc41c0e959d4de6124ce660d0d18c7f1c6423d672dee0a4dcf6","abstract_canon_sha256":"5af34457e41bc715348a91857cf52e3f282c682fed778c151a730203743016c7"},"schema_version":"1.0"},"canonical_sha256":"e7bf508bb6ceb6be6572fbcd76475765315536d79fa3e92864ae307124b0d2a9","source":{"kind":"arxiv","id":"2505.06284","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.06284","created_at":"2026-07-05T11:01:06Z"},{"alias_kind":"arxiv_version","alias_value":"2505.06284v1","created_at":"2026-07-05T11:01:06Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.06284","created_at":"2026-07-05T11:01:06Z"},{"alias_kind":"pith_short_12","alias_value":"467VBC5WZ23L","created_at":"2026-07-05T11:01:06Z"},{"alias_kind":"pith_short_16","alias_value":"467VBC5WZ23L4ZLS","created_at":"2026-07-05T11:01:06Z"},{"alias_kind":"pith_short_8","alias_value":"467VBC5W","created_at":"2026-07-05T11:01:06Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:467VBC5WZ23L4ZLS7PGXMR2XMU","target":"record","payload":{"canonical_record":{"source":{"id":"2505.06284","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-07T07:21:37Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"6297b2def7b58dc41c0e959d4de6124ce660d0d18c7f1c6423d672dee0a4dcf6","abstract_canon_sha256":"5af34457e41bc715348a91857cf52e3f282c682fed778c151a730203743016c7"},"schema_version":"1.0"},"canonical_sha256":"e7bf508bb6ceb6be6572fbcd76475765315536d79fa3e92864ae307124b0d2a9","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:01:06.614185Z","signature_b64":"RuYimNc+yDYkbeh2N6ZxsoGRj5F6TgcxTcPIqxGEfQkz0tbImWWEiMI3NqAqvhur879PXLksCWXjLFWKFXKfCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e7bf508bb6ceb6be6572fbcd76475765315536d79fa3e92864ae307124b0d2a9","last_reissued_at":"2026-07-05T11:01:06.613700Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:01:06.613700Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.06284","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:01:06Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3TO7EVOI2yRtsmXRQmOp7grB6avmzBp6rojZdPS1hWm+wcTOsJVEW22R78VVcSx6MPDh5o3Quklz0eUsUuOFDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-17T19:48:32.504052Z"},"content_sha256":"ba6aac1e058d78ce2b6ef79299580c58435452088273a48ab07805e8d003f3eb","schema_version":"1.0","event_id":"sha256:ba6aac1e058d78ce2b6ef79299580c58435452088273a48ab07805e8d003f3eb"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:467VBC5WZ23L4ZLS7PGXMR2XMU","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"DMRL: Data- and Model-aware Reward Learning for Data Extraction","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.LG","authors_text":"Ruoxi Cheng, Zhiqiang Wang","submitted_at":"2025-05-07T07:21:37Z","abstract_excerpt":"Large language models (LLMs) are inherently vulnerable to unintended privacy breaches. Consequently, systematic red-teaming research is essential for developing robust defense mechanisms. However, current data extraction methods suffer from several limitations: (1) rely on dataset duplicates (addressable via deduplication), (2) depend on prompt engineering (now countered by detection and defense), and (3) rely on random-search adversarial generation. To address these challenges, we propose DMRL, a Data- and Model-aware Reward Learning approach for data extraction. This technique leverages inve"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.06284","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/2505.06284/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:01:06Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8AbjBJKRuAziiFRMEF5+HX6sB0Zu7YnKT2OHtJPNOrdQagjIOnRLsXgnxB2NBJQluMO3aPX3Or1jnBjOqk92Aw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-17T19:48:32.504554Z"},"content_sha256":"95c68c04c7f5af2fca26a89da9fa019fbc1327de454b20529e5c11c8f14120ab","schema_version":"1.0","event_id":"sha256:95c68c04c7f5af2fca26a89da9fa019fbc1327de454b20529e5c11c8f14120ab"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/467VBC5WZ23L4ZLS7PGXMR2XMU/bundle.json","state_url":"https://pith.science/pith/467VBC5WZ23L4ZLS7PGXMR2XMU/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/467VBC5WZ23L4ZLS7PGXMR2XMU/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-17T19:48:32Z","links":{"resolver":"https://pith.science/pith/467VBC5WZ23L4ZLS7PGXMR2XMU","bundle":"https://pith.science/pith/467VBC5WZ23L4ZLS7PGXMR2XMU/bundle.json","state":"https://pith.science/pith/467VBC5WZ23L4ZLS7PGXMR2XMU/state.json","well_known_bundle":"https://pith.science/.well-known/pith/467VBC5WZ23L4ZLS7PGXMR2XMU/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:467VBC5WZ23L4ZLS7PGXMR2XMU","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"5af34457e41bc715348a91857cf52e3f282c682fed778c151a730203743016c7","cross_cats_sorted":["cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-07T07:21:37Z","title_canon_sha256":"6297b2def7b58dc41c0e959d4de6124ce660d0d18c7f1c6423d672dee0a4dcf6"},"schema_version":"1.0","source":{"id":"2505.06284","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.06284","created_at":"2026-07-05T11:01:06Z"},{"alias_kind":"arxiv_version","alias_value":"2505.06284v1","created_at":"2026-07-05T11:01:06Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.06284","created_at":"2026-07-05T11:01:06Z"},{"alias_kind":"pith_short_12","alias_value":"467VBC5WZ23L","created_at":"2026-07-05T11:01:06Z"},{"alias_kind":"pith_short_16","alias_value":"467VBC5WZ23L4ZLS","created_at":"2026-07-05T11:01:06Z"},{"alias_kind":"pith_short_8","alias_value":"467VBC5W","created_at":"2026-07-05T11:01:06Z"}],"graph_snapshots":[{"event_id":"sha256:95c68c04c7f5af2fca26a89da9fa019fbc1327de454b20529e5c11c8f14120ab","target":"graph","created_at":"2026-07-05T11:01:06Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2505.06284/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) are inherently vulnerable to unintended privacy breaches. Consequently, systematic red-teaming research is essential for developing robust defense mechanisms. However, current data extraction methods suffer from several limitations: (1) rely on dataset duplicates (addressable via deduplication), (2) depend on prompt engineering (now countered by detection and defense), and (3) rely on random-search adversarial generation. To address these challenges, we propose DMRL, a Data- and Model-aware Reward Learning approach for data extraction. This technique leverages inve","authors_text":"Ruoxi Cheng, Zhiqiang Wang","cross_cats":["cs.CR"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-07T07:21:37Z","title":"DMRL: Data- and Model-aware Reward Learning for Data Extraction"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.06284","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:ba6aac1e058d78ce2b6ef79299580c58435452088273a48ab07805e8d003f3eb","target":"record","created_at":"2026-07-05T11:01:06Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"5af34457e41bc715348a91857cf52e3f282c682fed778c151a730203743016c7","cross_cats_sorted":["cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-07T07:21:37Z","title_canon_sha256":"6297b2def7b58dc41c0e959d4de6124ce660d0d18c7f1c6423d672dee0a4dcf6"},"schema_version":"1.0","source":{"id":"2505.06284","kind":"arxiv","version":1}},"canonical_sha256":"e7bf508bb6ceb6be6572fbcd76475765315536d79fa3e92864ae307124b0d2a9","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e7bf508bb6ceb6be6572fbcd76475765315536d79fa3e92864ae307124b0d2a9","first_computed_at":"2026-07-05T11:01:06.613700Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:01:06.613700Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"RuYimNc+yDYkbeh2N6ZxsoGRj5F6TgcxTcPIqxGEfQkz0tbImWWEiMI3NqAqvhur879PXLksCWXjLFWKFXKfCg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:01:06.614185Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.06284","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:ba6aac1e058d78ce2b6ef79299580c58435452088273a48ab07805e8d003f3eb","sha256:95c68c04c7f5af2fca26a89da9fa019fbc1327de454b20529e5c11c8f14120ab"],"state_sha256":"cc08e671666d51cd632301f5b354acf34dacb1ca9a7bfa840b359becb7f43b92"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7otDmISVabuXgxuFjREx4GSZDdJVx8GkG8p7pvxB33fyuLwSmiX78QbILQD8vf88POHOJXAZsD38Dg5Tg0tJDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-17T19:48:32.508230Z","bundle_sha256":"de5a5ba129024b3d1246e1abc502c03cc14957311fbaf39dd42fee25d044e10b"}}