{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:4EUURVWS5XN2ZIFZ3UTB66JPHK","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":"6ae3d2f806e8310b808808e7cbf430dca23e9f052a0e3100f97272aff360b8dd","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-02-20T12:07:40Z","title_canon_sha256":"46901bb89ef113bb0095eee0bfefc1cb8f58a58a681ecac9e2dd57ba1b0805de"},"schema_version":"1.0","source":{"id":"2502.14486","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.14486","created_at":"2026-07-05T10:17:27Z"},{"alias_kind":"arxiv_version","alias_value":"2502.14486v1","created_at":"2026-07-05T10:17:27Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.14486","created_at":"2026-07-05T10:17:27Z"},{"alias_kind":"pith_short_12","alias_value":"4EUURVWS5XN2","created_at":"2026-07-05T10:17:27Z"},{"alias_kind":"pith_short_16","alias_value":"4EUURVWS5XN2ZIFZ","created_at":"2026-07-05T10:17:27Z"},{"alias_kind":"pith_short_8","alias_value":"4EUURVWS","created_at":"2026-07-05T10:17:27Z"}],"graph_snapshots":[{"event_id":"sha256:37bce153d2ae46768aa21fe6b226c4ef5e82354717306127511442750c3963ea","target":"graph","created_at":"2026-07-05T10:17:27Z","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/2502.14486/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Jailbreak attacks, where harmful prompts bypass generative models' built-in safety, raise serious concerns about model vulnerability. While many defense methods have been proposed, the trade-offs between safety and helpfulness, and their application to Large Vision-Language Models (LVLMs), are not well understood. This paper systematically examines jailbreak defenses by reframing the standard generation task as a binary classification problem to assess model refusal tendencies for both harmful and benign queries. We identify two key defense mechanisms: safety shift, which increases refusal rat","authors_text":"Shujun Liu, Siyuan Wang, Xuanjing Huang, Yuhang Lai, Zhongyu Wei, Zhuohang Long","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-02-20T12:07:40Z","title":"How Jailbreak Defenses Work and Ensemble? A Mechanistic Investigation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.14486","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":"sha256:020eb18d041fad21d2dc1fbd1e6424645231c245eaeecb8c188696c9544c2df6","target":"record","created_at":"2026-07-05T10:17:27Z","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":"6ae3d2f806e8310b808808e7cbf430dca23e9f052a0e3100f97272aff360b8dd","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-02-20T12:07:40Z","title_canon_sha256":"46901bb89ef113bb0095eee0bfefc1cb8f58a58a681ecac9e2dd57ba1b0805de"},"schema_version":"1.0","source":{"id":"2502.14486","kind":"arxiv","version":1}},"canonical_sha256":"e12948d6d2eddbaca0b9dd261f792f3ab20d04f9c128110212b98bcc8b7d703b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e12948d6d2eddbaca0b9dd261f792f3ab20d04f9c128110212b98bcc8b7d703b","first_computed_at":"2026-07-05T10:17:27.392514Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:17:27.392514Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"KZJB3p6Li8vaaxT/INSBIb5Ndi0RBxeYEA2mpzj5Dfj34Z8h29pZEaVFpz4aba7YU1C0J+4DyDlwx70vryNrDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T10:17:27.393019Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.14486","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:020eb18d041fad21d2dc1fbd1e6424645231c245eaeecb8c188696c9544c2df6","sha256:37bce153d2ae46768aa21fe6b226c4ef5e82354717306127511442750c3963ea"],"state_sha256":"e219b2c8a531c96bd7b184ebc0f3c45c0f5cf03763fdd0265c889755aeab7497"}