{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:SCE6ZV3C6JMA6DUAEOOGBDIV4Y","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":"b286d7453e825fcd8bb76bda86d42660543e2e3d2cfc172e93c095cb46df434b","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-19T07:03:05Z","title_canon_sha256":"9ffc85703d670e0ab8a55731e53aaf8972996e9fdb3523d311a46912655113dc"},"schema_version":"1.0","source":{"id":"2506.16078","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.16078","created_at":"2026-07-05T11:24:30Z"},{"alias_kind":"arxiv_version","alias_value":"2506.16078v1","created_at":"2026-07-05T11:24:30Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.16078","created_at":"2026-07-05T11:24:30Z"},{"alias_kind":"pith_short_12","alias_value":"SCE6ZV3C6JMA","created_at":"2026-07-05T11:24:30Z"},{"alias_kind":"pith_short_16","alias_value":"SCE6ZV3C6JMA6DUA","created_at":"2026-07-05T11:24:30Z"},{"alias_kind":"pith_short_8","alias_value":"SCE6ZV3C","created_at":"2026-07-05T11:24:30Z"}],"graph_snapshots":[{"event_id":"sha256:9ff47c205f4d28119c6b9c96dc4e675032439f716286b12e86cbd669c6e6d851","target":"graph","created_at":"2026-07-05T11:24:30Z","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/2506.16078/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Safety alignment is a key requirement for building reliable Artificial General Intelligence. Despite significant advances in safety alignment, we observe that minor latent shifts can still trigger unsafe responses in aligned models. We argue that this stems from the shallow nature of existing alignment methods, which focus on surface-level refusal behaviors without sufficiently altering internal representations. Consequently, small shifts in hidden activations can re-trigger harmful behaviors embedded in the latent space. To explore the robustness of safety alignment to latent perturbations, w","authors_text":"Jie Li, Kexin Huang, Tianle Gu, Yang Yao, Yan Teng, Yingchun Wang, Yixu Wang, Yuanqi Yao, Yujiu Yang, Zongqi Wang","cross_cats":["cs.AI","cs.CL","cs.CR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-19T07:03:05Z","title":"Probing the Robustness of Large Language Models Safety to Latent Perturbations"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.16078","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":"sha256:8aeeafccaf2101f1f009df7deb4efb12e5c0f9924e4b95ac2c35a052c3e99ec3","target":"record","created_at":"2026-07-05T11:24:30Z","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":"b286d7453e825fcd8bb76bda86d42660543e2e3d2cfc172e93c095cb46df434b","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-19T07:03:05Z","title_canon_sha256":"9ffc85703d670e0ab8a55731e53aaf8972996e9fdb3523d311a46912655113dc"},"schema_version":"1.0","source":{"id":"2506.16078","kind":"arxiv","version":1}},"canonical_sha256":"9089ecd762f2580f0e80239c608d15e6377257f1d545b55355d65e18894cca82","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9089ecd762f2580f0e80239c608d15e6377257f1d545b55355d65e18894cca82","first_computed_at":"2026-07-05T11:24:30.060023Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:24:30.060023Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"nJNVtEiTz9FvbHtntcFyjs4mD6d6TLvWxETaDf2FxeMNG0C2pTOtzwS0tuFT1lMs5npbZRBbP0RvTCUVL+JUAA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:24:30.060622Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.16078","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8aeeafccaf2101f1f009df7deb4efb12e5c0f9924e4b95ac2c35a052c3e99ec3","sha256:9ff47c205f4d28119c6b9c96dc4e675032439f716286b12e86cbd669c6e6d851"],"state_sha256":"0252b0850a1a49c3444524e655a225ad09b65b026895a6891f306aeed6765032"}