{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:4EUURVWS5XN2ZIFZ3UTB66JPHK","short_pith_number":"pith:4EUURVWS","canonical_record":{"source":{"id":"2502.14486","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-02-20T12:07:40Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"46901bb89ef113bb0095eee0bfefc1cb8f58a58a681ecac9e2dd57ba1b0805de","abstract_canon_sha256":"6ae3d2f806e8310b808808e7cbf430dca23e9f052a0e3100f97272aff360b8dd"},"schema_version":"1.0"},"canonical_sha256":"e12948d6d2eddbaca0b9dd261f792f3ab20d04f9c128110212b98bcc8b7d703b","source":{"kind":"arxiv","id":"2502.14486","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"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:4EUURVWS5XN2ZIFZ3UTB66JPHK","target":"record","payload":{"canonical_record":{"source":{"id":"2502.14486","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-02-20T12:07:40Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"46901bb89ef113bb0095eee0bfefc1cb8f58a58a681ecac9e2dd57ba1b0805de","abstract_canon_sha256":"6ae3d2f806e8310b808808e7cbf430dca23e9f052a0e3100f97272aff360b8dd"},"schema_version":"1.0"},"canonical_sha256":"e12948d6d2eddbaca0b9dd261f792f3ab20d04f9c128110212b98bcc8b7d703b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:17:27.393019Z","signature_b64":"KZJB3p6Li8vaaxT/INSBIb5Ndi0RBxeYEA2mpzj5Dfj34Z8h29pZEaVFpz4aba7YU1C0J+4DyDlwx70vryNrDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e12948d6d2eddbaca0b9dd261f792f3ab20d04f9c128110212b98bcc8b7d703b","last_reissued_at":"2026-07-05T10:17:27.392514Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:17:27.392514Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2502.14486","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":"2026-07-05T10:17:27Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"WSSU/lZl51m5vM0rhpAf8khOOdGxvj9kfxdQcJW5f2bQbrjeXT+fw1LuaVRF72Fs2xUalWVyucZBLIkuM1O0Dw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T09:05:09.093533Z"},"content_sha256":"020eb18d041fad21d2dc1fbd1e6424645231c245eaeecb8c188696c9544c2df6","schema_version":"1.0","event_id":"sha256:020eb18d041fad21d2dc1fbd1e6424645231c245eaeecb8c188696c9544c2df6"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:4EUURVWS5XN2ZIFZ3UTB66JPHK","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"How Jailbreak Defenses Work and Ensemble? A Mechanistic Investigation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Shujun Liu, Siyuan Wang, Xuanjing Huang, Yuhang Lai, Zhongyu Wei, Zhuohang Long","submitted_at":"2025-02-20T12:07:40Z","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"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.14486","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/2502.14486/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":"2026-07-05T10:17:27Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MOsDKswCSlm9Lfe6S7gCQIPL8Jf88cpT8mmKLIIRNwZNYJQuXritCvNGOEe4OF2Rv6KPIF7AvShVkLN7dWpsDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T09:05:09.094040Z"},"content_sha256":"37bce153d2ae46768aa21fe6b226c4ef5e82354717306127511442750c3963ea","schema_version":"1.0","event_id":"sha256:37bce153d2ae46768aa21fe6b226c4ef5e82354717306127511442750c3963ea"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/4EUURVWS5XN2ZIFZ3UTB66JPHK/bundle.json","state_url":"https://pith.science/pith/4EUURVWS5XN2ZIFZ3UTB66JPHK/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/4EUURVWS5XN2ZIFZ3UTB66JPHK/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":"2026-08-06T09:05:09Z","links":{"resolver":"https://pith.science/pith/4EUURVWS5XN2ZIFZ3UTB66JPHK","bundle":"https://pith.science/pith/4EUURVWS5XN2ZIFZ3UTB66JPHK/bundle.json","state":"https://pith.science/pith/4EUURVWS5XN2ZIFZ3UTB66JPHK/state.json","well_known_bundle":"https://pith.science/.well-known/pith/4EUURVWS5XN2ZIFZ3UTB66JPHK/bundle.json"},"state":{"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"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"wEmiRFshXdLAF4t/A/nnxjPWkt/yrdvSUPQdHdH7ZZ3u/EKJBjUHR+uExhpqHfFgiOnIldxyoqr8Aq95H53/BQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T09:05:09.101323Z","bundle_sha256":"83ca566de065be163ed825873946deb7bef4f486e6ad860ffe52fec9dcbd1727"}}