{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:NGXLJBOE3DM3S7WD4FWMQY7UQ3","short_pith_number":"pith:NGXLJBOE","canonical_record":{"source":{"id":"2507.02332","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-03T05:50:50Z","cross_cats_sorted":[],"title_canon_sha256":"dad75e616f4cb1563fe4f2019cb533bc11636278d11eaf2691562d0376d50e97","abstract_canon_sha256":"a64cfc0c4d2192e4512c0a6f6777ca7fe3c3c3883afc35e478b0cf6018f46e46"},"schema_version":"1.0"},"canonical_sha256":"69aeb485c4d8d9b97ec3e16cc863f486e5dedf7deb87d62b2a759ac2a790ed37","source":{"kind":"arxiv","id":"2507.02332","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.02332","created_at":"2026-07-05T11:55:41Z"},{"alias_kind":"arxiv_version","alias_value":"2507.02332v2","created_at":"2026-07-05T11:55:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.02332","created_at":"2026-07-05T11:55:41Z"},{"alias_kind":"pith_short_12","alias_value":"NGXLJBOE3DM3","created_at":"2026-07-05T11:55:41Z"},{"alias_kind":"pith_short_16","alias_value":"NGXLJBOE3DM3S7WD","created_at":"2026-07-05T11:55:41Z"},{"alias_kind":"pith_short_8","alias_value":"NGXLJBOE","created_at":"2026-07-05T11:55:41Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:NGXLJBOE3DM3S7WD4FWMQY7UQ3","target":"record","payload":{"canonical_record":{"source":{"id":"2507.02332","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-03T05:50:50Z","cross_cats_sorted":[],"title_canon_sha256":"dad75e616f4cb1563fe4f2019cb533bc11636278d11eaf2691562d0376d50e97","abstract_canon_sha256":"a64cfc0c4d2192e4512c0a6f6777ca7fe3c3c3883afc35e478b0cf6018f46e46"},"schema_version":"1.0"},"canonical_sha256":"69aeb485c4d8d9b97ec3e16cc863f486e5dedf7deb87d62b2a759ac2a790ed37","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:55:41.108075Z","signature_b64":"63MOM3KgAx5jNiKdkn1NRgR3O6329f2AkEGJGFj58Kmc8N7sLmWoZylNyxe0ZvRwm3zNUczsmNPy/YkicxD/Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"69aeb485c4d8d9b97ec3e16cc863f486e5dedf7deb87d62b2a759ac2a790ed37","last_reissued_at":"2026-07-05T11:55:41.107692Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:55:41.107692Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2507.02332","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:55:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SwnozWBDY2FY/GOrj9WFxKc9NFv9Zewuvb4e2Kq9MP+lkzPXFaxgGWsXfEvZa4LhrjGhlP1glr1H3t3xVCQGCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T19:07:24.325936Z"},"content_sha256":"a1f84fc78e8e43fbf2c1ea884ff658318dcf2e867e2094d49fc2b29886bfece2","schema_version":"1.0","event_id":"sha256:a1f84fc78e8e43fbf2c1ea884ff658318dcf2e867e2094d49fc2b29886bfece2"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:NGXLJBOE3DM3S7WD4FWMQY7UQ3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"PII Jailbreaking in LLMs via Activation Steering Reveals Personal Information Leakage","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Dmitrii Usynin, Krishna Kanth Nakka, Xuebing Zhou, Xue Jiang","submitted_at":"2025-07-03T05:50:50Z","abstract_excerpt":"This paper investigates privacy jailbreaking in LLMs via steering, focusing on whether manipulating activations can bypass LLM alignment and alter response behaviors to privacy related queries (e.g., a certain public figure's sexual orientation). We begin by identifying attention heads predictive of refusal behavior for private attributes (e.g., sexual orientation) using lightweight linear probes trained with privacy evaluator labels. Next, we steer the activations of a small subset of these attention heads guided by the trained probes to induce the model to generate non-refusal responses. Our"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.02332","kind":"arxiv","version":2},"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/2507.02332/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":"2026-07-05T11:55:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"GqBYUQi7Y+S4htwlhIOULpn3ESsZPOY8wI1QpmLwfJHgHVv4tntdB3KAyK3y6YnvpQlljpqRf3SEkPjt5V5hDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T19:07:24.326310Z"},"content_sha256":"2c2eaf6ff12ebfa793db604cbabb52833b290a81067a2b107bc53944a6ee2d61","schema_version":"1.0","event_id":"sha256:2c2eaf6ff12ebfa793db604cbabb52833b290a81067a2b107bc53944a6ee2d61"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/NGXLJBOE3DM3S7WD4FWMQY7UQ3/bundle.json","state_url":"https://pith.science/pith/NGXLJBOE3DM3S7WD4FWMQY7UQ3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/NGXLJBOE3DM3S7WD4FWMQY7UQ3/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":"2026-07-06T19:07:24Z","links":{"resolver":"https://pith.science/pith/NGXLJBOE3DM3S7WD4FWMQY7UQ3","bundle":"https://pith.science/pith/NGXLJBOE3DM3S7WD4FWMQY7UQ3/bundle.json","state":"https://pith.science/pith/NGXLJBOE3DM3S7WD4FWMQY7UQ3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/NGXLJBOE3DM3S7WD4FWMQY7UQ3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:NGXLJBOE3DM3S7WD4FWMQY7UQ3","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":"a64cfc0c4d2192e4512c0a6f6777ca7fe3c3c3883afc35e478b0cf6018f46e46","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-03T05:50:50Z","title_canon_sha256":"dad75e616f4cb1563fe4f2019cb533bc11636278d11eaf2691562d0376d50e97"},"schema_version":"1.0","source":{"id":"2507.02332","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.02332","created_at":"2026-07-05T11:55:41Z"},{"alias_kind":"arxiv_version","alias_value":"2507.02332v2","created_at":"2026-07-05T11:55:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.02332","created_at":"2026-07-05T11:55:41Z"},{"alias_kind":"pith_short_12","alias_value":"NGXLJBOE3DM3","created_at":"2026-07-05T11:55:41Z"},{"alias_kind":"pith_short_16","alias_value":"NGXLJBOE3DM3S7WD","created_at":"2026-07-05T11:55:41Z"},{"alias_kind":"pith_short_8","alias_value":"NGXLJBOE","created_at":"2026-07-05T11:55:41Z"}],"graph_snapshots":[{"event_id":"sha256:2c2eaf6ff12ebfa793db604cbabb52833b290a81067a2b107bc53944a6ee2d61","target":"graph","created_at":"2026-07-05T11:55:41Z","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/2507.02332/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"This paper investigates privacy jailbreaking in LLMs via steering, focusing on whether manipulating activations can bypass LLM alignment and alter response behaviors to privacy related queries (e.g., a certain public figure's sexual orientation). We begin by identifying attention heads predictive of refusal behavior for private attributes (e.g., sexual orientation) using lightweight linear probes trained with privacy evaluator labels. Next, we steer the activations of a small subset of these attention heads guided by the trained probes to induce the model to generate non-refusal responses. Our","authors_text":"Dmitrii Usynin, Krishna Kanth Nakka, Xuebing Zhou, Xue Jiang","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-03T05:50:50Z","title":"PII Jailbreaking in LLMs via Activation Steering Reveals Personal Information Leakage"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.02332","kind":"arxiv","version":2},"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:a1f84fc78e8e43fbf2c1ea884ff658318dcf2e867e2094d49fc2b29886bfece2","target":"record","created_at":"2026-07-05T11:55:41Z","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":"a64cfc0c4d2192e4512c0a6f6777ca7fe3c3c3883afc35e478b0cf6018f46e46","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-03T05:50:50Z","title_canon_sha256":"dad75e616f4cb1563fe4f2019cb533bc11636278d11eaf2691562d0376d50e97"},"schema_version":"1.0","source":{"id":"2507.02332","kind":"arxiv","version":2}},"canonical_sha256":"69aeb485c4d8d9b97ec3e16cc863f486e5dedf7deb87d62b2a759ac2a790ed37","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"69aeb485c4d8d9b97ec3e16cc863f486e5dedf7deb87d62b2a759ac2a790ed37","first_computed_at":"2026-07-05T11:55:41.107692Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:55:41.107692Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"63MOM3KgAx5jNiKdkn1NRgR3O6329f2AkEGJGFj58Kmc8N7sLmWoZylNyxe0ZvRwm3zNUczsmNPy/YkicxD/Ag==","signature_status":"signed_v1","signed_at":"2026-07-05T11:55:41.108075Z","signed_message":"canonical_sha256_bytes"},"source_id":"2507.02332","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a1f84fc78e8e43fbf2c1ea884ff658318dcf2e867e2094d49fc2b29886bfece2","sha256:2c2eaf6ff12ebfa793db604cbabb52833b290a81067a2b107bc53944a6ee2d61"],"state_sha256":"198dd050786aa5b11aab30580f81c30b20e6ebf6de4a913dffa4a7eecb43a33e"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"jeNsHkAQJtTJYLGC2GUQv9hXNcxxHgs9wWmIC7nz+0n+v+fY3W9u34RjB3N5DoA9gvdoJwTukegX+nYqEuuKDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-06T19:07:24.328174Z","bundle_sha256":"b64f20ad4b3c42f815f6efc9d36ebe1d7711879a9464a9af4adb6a1d324434f0"}}