{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:3OV3ESHUDJ4ULZ34EXOM7VZWRV","short_pith_number":"pith:3OV3ESHU","canonical_record":{"source":{"id":"2401.17263","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-01-30T18:56:08Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CV"],"title_canon_sha256":"af54f738b7d646b9e17909e297213be56335b007132091a802c581242091af41","abstract_canon_sha256":"5c07699b0c5a295b978b7491959f8dab6bba40e63a903b793d85756919bf7adc"},"schema_version":"1.0"},"canonical_sha256":"dbabb248f41a7945e77c25dccfd7368d7cfd03eb25bc81cd066e4f72dc12bc52","source":{"kind":"arxiv","id":"2401.17263","version":5},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.17263","created_at":"2026-07-05T09:32:41Z"},{"alias_kind":"arxiv_version","alias_value":"2401.17263v5","created_at":"2026-07-05T09:32:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.17263","created_at":"2026-07-05T09:32:41Z"},{"alias_kind":"pith_short_12","alias_value":"3OV3ESHUDJ4U","created_at":"2026-07-05T09:32:41Z"},{"alias_kind":"pith_short_16","alias_value":"3OV3ESHUDJ4ULZ34","created_at":"2026-07-05T09:32:41Z"},{"alias_kind":"pith_short_8","alias_value":"3OV3ESHU","created_at":"2026-07-05T09:32:41Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:3OV3ESHUDJ4ULZ34EXOM7VZWRV","target":"record","payload":{"canonical_record":{"source":{"id":"2401.17263","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-01-30T18:56:08Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CV"],"title_canon_sha256":"af54f738b7d646b9e17909e297213be56335b007132091a802c581242091af41","abstract_canon_sha256":"5c07699b0c5a295b978b7491959f8dab6bba40e63a903b793d85756919bf7adc"},"schema_version":"1.0"},"canonical_sha256":"dbabb248f41a7945e77c25dccfd7368d7cfd03eb25bc81cd066e4f72dc12bc52","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:32:41.854597Z","signature_b64":"HdObim88BOY7J27kZb0pk9JNGxqOhvmp1HCKPg5amtA66DunGUGJB1iFxy4XcF24Ltj6dV/tFcUM4Ja5eAqlAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dbabb248f41a7945e77c25dccfd7368d7cfd03eb25bc81cd066e4f72dc12bc52","last_reissued_at":"2026-07-05T09:32:41.854090Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:32:41.854090Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2401.17263","source_version":5,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:32:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"jmBu/OrAunuWLpQ6N+7MUWikpbDTO4jlSAEYIjGiaAPvcXbKwQFxFICwRTmVxpH5n6xSb4JlfLh24o/lxC4GBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T22:56:10.750045Z"},"content_sha256":"4fc92991ac0aae448b9ab9747423cf29d27311cde804572265dd594288a8d73b","schema_version":"1.0","event_id":"sha256:4fc92991ac0aae448b9ab9747423cf29d27311cde804572265dd594288a8d73b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:3OV3ESHUDJ4ULZ34EXOM7VZWRV","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Robust Prompt Optimization for Defending Language Models Against Jailbreaking Attacks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.CV"],"primary_cat":"cs.LG","authors_text":"Andy Zhou, Bo Li, Haohan Wang","submitted_at":"2024-01-30T18:56:08Z","abstract_excerpt":"Despite advances in AI alignment, large language models (LLMs) remain vulnerable to adversarial attacks or jailbreaking, in which adversaries can modify prompts to induce unwanted behavior. While some defenses have been proposed, they have not been adapted to newly proposed attacks and more challenging threat models. To address this, we propose an optimization-based objective for defending LLMs against jailbreaking attacks and an algorithm, Robust Prompt Optimization (RPO) to create robust system-level defenses. Our approach directly incorporates the adversary into the defensive objective and "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.17263","kind":"arxiv","version":5},"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/2401.17263/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":"2026-07-05T09:32:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SFYD1v6xJPUAaHJm9W59Q5YiqupT5R/OrVaKRr64UnMpBr6gbBp+hfpgWFuaNuV6izQvRlVWPZBIAEnLP6xyDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-05T22:56:10.750556Z"},"content_sha256":"6ebad1c58928c142190e76261cadf17a7f4d5d0259938c83665a73e131daf302","schema_version":"1.0","event_id":"sha256:6ebad1c58928c142190e76261cadf17a7f4d5d0259938c83665a73e131daf302"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/3OV3ESHUDJ4ULZ34EXOM7VZWRV/bundle.json","state_url":"https://pith.science/pith/3OV3ESHUDJ4ULZ34EXOM7VZWRV/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/3OV3ESHUDJ4ULZ34EXOM7VZWRV/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":"2026-08-05T22:56:10Z","links":{"resolver":"https://pith.science/pith/3OV3ESHUDJ4ULZ34EXOM7VZWRV","bundle":"https://pith.science/pith/3OV3ESHUDJ4ULZ34EXOM7VZWRV/bundle.json","state":"https://pith.science/pith/3OV3ESHUDJ4ULZ34EXOM7VZWRV/state.json","well_known_bundle":"https://pith.science/.well-known/pith/3OV3ESHUDJ4ULZ34EXOM7VZWRV/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:3OV3ESHUDJ4ULZ34EXOM7VZWRV","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":"5c07699b0c5a295b978b7491959f8dab6bba40e63a903b793d85756919bf7adc","cross_cats_sorted":["cs.AI","cs.CL","cs.CV"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-01-30T18:56:08Z","title_canon_sha256":"af54f738b7d646b9e17909e297213be56335b007132091a802c581242091af41"},"schema_version":"1.0","source":{"id":"2401.17263","kind":"arxiv","version":5}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.17263","created_at":"2026-07-05T09:32:41Z"},{"alias_kind":"arxiv_version","alias_value":"2401.17263v5","created_at":"2026-07-05T09:32:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.17263","created_at":"2026-07-05T09:32:41Z"},{"alias_kind":"pith_short_12","alias_value":"3OV3ESHUDJ4U","created_at":"2026-07-05T09:32:41Z"},{"alias_kind":"pith_short_16","alias_value":"3OV3ESHUDJ4ULZ34","created_at":"2026-07-05T09:32:41Z"},{"alias_kind":"pith_short_8","alias_value":"3OV3ESHU","created_at":"2026-07-05T09:32:41Z"}],"graph_snapshots":[{"event_id":"sha256:6ebad1c58928c142190e76261cadf17a7f4d5d0259938c83665a73e131daf302","target":"graph","created_at":"2026-07-05T09:32:41Z","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/2401.17263/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Despite advances in AI alignment, large language models (LLMs) remain vulnerable to adversarial attacks or jailbreaking, in which adversaries can modify prompts to induce unwanted behavior. While some defenses have been proposed, they have not been adapted to newly proposed attacks and more challenging threat models. To address this, we propose an optimization-based objective for defending LLMs against jailbreaking attacks and an algorithm, Robust Prompt Optimization (RPO) to create robust system-level defenses. Our approach directly incorporates the adversary into the defensive objective and ","authors_text":"Andy Zhou, Bo Li, Haohan Wang","cross_cats":["cs.AI","cs.CL","cs.CV"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-01-30T18:56:08Z","title":"Robust Prompt Optimization for Defending Language Models Against Jailbreaking Attacks"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.17263","kind":"arxiv","version":5},"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":"sha256:4fc92991ac0aae448b9ab9747423cf29d27311cde804572265dd594288a8d73b","target":"record","created_at":"2026-07-05T09:32:41Z","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":"5c07699b0c5a295b978b7491959f8dab6bba40e63a903b793d85756919bf7adc","cross_cats_sorted":["cs.AI","cs.CL","cs.CV"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-01-30T18:56:08Z","title_canon_sha256":"af54f738b7d646b9e17909e297213be56335b007132091a802c581242091af41"},"schema_version":"1.0","source":{"id":"2401.17263","kind":"arxiv","version":5}},"canonical_sha256":"dbabb248f41a7945e77c25dccfd7368d7cfd03eb25bc81cd066e4f72dc12bc52","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"dbabb248f41a7945e77c25dccfd7368d7cfd03eb25bc81cd066e4f72dc12bc52","first_computed_at":"2026-07-05T09:32:41.854090Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:32:41.854090Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"HdObim88BOY7J27kZb0pk9JNGxqOhvmp1HCKPg5amtA66DunGUGJB1iFxy4XcF24Ltj6dV/tFcUM4Ja5eAqlAw==","signature_status":"signed_v1","signed_at":"2026-07-05T09:32:41.854597Z","signed_message":"canonical_sha256_bytes"},"source_id":"2401.17263","source_kind":"arxiv","source_version":5}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4fc92991ac0aae448b9ab9747423cf29d27311cde804572265dd594288a8d73b","sha256:6ebad1c58928c142190e76261cadf17a7f4d5d0259938c83665a73e131daf302"],"state_sha256":"ee7aaa0572f6493dd5d39af7af3290ff8b9b8fff454db5ff49540e49d5baf819"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"z4GUaKC1yMG366TBdNvpT5hb22DgAjs0MkGIFPT2mRFyv+Nz1L7RayGpq3MuKSAlraPz4tkTh61izfdZDdO3AQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-05T22:56:10.754488Z","bundle_sha256":"834cf42ae2f6c4ad8b47ec573c14505a406151d7c204130ee7f41f85270a6698"}}