{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:MCPJVVI5ACFGUK7EXMFNBCMC43","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":"a4932c8319b17028d9cfab0decebb5e0bff48f12f60bf55acd88edf5bd9d75c2","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-05-26T17:40:40Z","title_canon_sha256":"5a78d58d3d8402c8c3e3b29712e8f12889efb226cd09e37d40540d9122badf19"},"schema_version":"1.0","source":{"id":"2505.20259","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.20259","created_at":"2026-07-05T11:09:55Z"},{"alias_kind":"arxiv_version","alias_value":"2505.20259v1","created_at":"2026-07-05T11:09:55Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.20259","created_at":"2026-07-05T11:09:55Z"},{"alias_kind":"pith_short_12","alias_value":"MCPJVVI5ACFG","created_at":"2026-07-05T11:09:55Z"},{"alias_kind":"pith_short_16","alias_value":"MCPJVVI5ACFGUK7E","created_at":"2026-07-05T11:09:55Z"},{"alias_kind":"pith_short_8","alias_value":"MCPJVVI5","created_at":"2026-07-05T11:09:55Z"}],"graph_snapshots":[{"event_id":"sha256:ae64ae26534dc2a41c02f18e57946a42967abd45d8a472646e7b899181d6bb94","target":"graph","created_at":"2026-07-05T11:09:55Z","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/2505.20259/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"LLMs have made impressive progress, but their growing capabilities also expose them to highly flexible jailbreaking attacks designed to bypass safety alignment. While many existing defenses focus on known types of attacks, it is more critical to prepare LLMs for unseen attacks that may arise during deployment. To address this, we propose a lifelong safety alignment framework that enables LLMs to continuously adapt to new and evolving jailbreaking strategies. Our framework introduces a competitive setup between two components: a Meta-Attacker, trained to actively discover novel jailbreaking str","authors_text":"Chao Du, Haoyu Wang, Min Lin, Tianyu Pang, Xueqian Wang, Yifei Zhao, Zeyu Qin","cross_cats":["cs.AI","cs.CL","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-05-26T17:40:40Z","title":"Lifelong Safety Alignment for Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.20259","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:acc3ca26325f3539006b768edb021033deff39cef3e4dfd1dbf8e867a3bd9e16","target":"record","created_at":"2026-07-05T11:09:55Z","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":"a4932c8319b17028d9cfab0decebb5e0bff48f12f60bf55acd88edf5bd9d75c2","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-05-26T17:40:40Z","title_canon_sha256":"5a78d58d3d8402c8c3e3b29712e8f12889efb226cd09e37d40540d9122badf19"},"schema_version":"1.0","source":{"id":"2505.20259","kind":"arxiv","version":1}},"canonical_sha256":"609e9ad51d008a6a2be4bb0ad08982e6d13c829b47b9da876a1b7cbddb0133fa","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"609e9ad51d008a6a2be4bb0ad08982e6d13c829b47b9da876a1b7cbddb0133fa","first_computed_at":"2026-07-05T11:09:55.282405Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:09:55.282405Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"RcJ+g6kEetCABtDC8dVF9V25bcS9OwyFSBiCyxk9zOqmspTcGrKSCvVh0N0P/MI/2ZYJJGM+FaYYGREOZoXuAg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:09:55.282880Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.20259","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:acc3ca26325f3539006b768edb021033deff39cef3e4dfd1dbf8e867a3bd9e16","sha256:ae64ae26534dc2a41c02f18e57946a42967abd45d8a472646e7b899181d6bb94"],"state_sha256":"24680d5deaba2cad2c2f7cf03be00d42ac2e70439ad74e237c9df5ed2d77249d"}