{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:UJJKWNSGMYPRXMJD7V3AVYQBE6","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":"3229637d1ce9816cb8c5b1a4684529c3d4edc5d2d8113cd6e5f430395161802b","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-06T03:41:44Z","title_canon_sha256":"588667773725635bd06edd39097b303dbab958554258567baf5326c30ec0c322"},"schema_version":"1.0","source":{"id":"2606.07963","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.07963","created_at":"2026-06-09T01:04:56Z"},{"alias_kind":"arxiv_version","alias_value":"2606.07963v1","created_at":"2026-06-09T01:04:56Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.07963","created_at":"2026-06-09T01:04:56Z"},{"alias_kind":"pith_short_12","alias_value":"UJJKWNSGMYPR","created_at":"2026-06-09T01:04:56Z"},{"alias_kind":"pith_short_16","alias_value":"UJJKWNSGMYPRXMJD","created_at":"2026-06-09T01:04:56Z"},{"alias_kind":"pith_short_8","alias_value":"UJJKWNSG","created_at":"2026-06-09T01:04:56Z"}],"graph_snapshots":[{"event_id":"sha256:6d035922c1d2e647d93455b6989208e6f4e4397d5672de1b79c4573a53625154","target":"graph","created_at":"2026-06-09T01:04:56Z","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/2606.07963/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Backdoor attacks in large language models (LLMs) are often treated as isolated trigger-response failures, motivating defenses tailored to specific triggers or behaviors. We show this view is incomplete. Across diverse backdoor behaviors, we identify a shared latent mechanism that can be detected, causally controlled, and suppressed. Using sparse autoencoders (SAEs) on residual-stream activations, we find a small set of latent features consistently activated across jailbreaking, refusal manipulation, password-locking, bias induction, sentiment misclassification, and country-conditioned harmful ","authors_text":"Aly M. Kassem, Buddhika Laknath Semage, Golnoosh Farnadi, Negar Rostamzadeh, Omar Mahmoud, Santu Rana, Thommen George Karimpanal","cross_cats":["cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-06T03:41:44Z","title":"Shared Latent Structures Enable Unified Backdoor Detection and Mitigation in LLMs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.07963","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:a48b4240944b9b33825775f16e34b92a0c9d24c3536aa635a9bbf26dbae78e0a","target":"record","created_at":"2026-06-09T01:04:56Z","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":"3229637d1ce9816cb8c5b1a4684529c3d4edc5d2d8113cd6e5f430395161802b","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-06T03:41:44Z","title_canon_sha256":"588667773725635bd06edd39097b303dbab958554258567baf5326c30ec0c322"},"schema_version":"1.0","source":{"id":"2606.07963","kind":"arxiv","version":1}},"canonical_sha256":"a252ab3646661f1bb123fd760ae20127bd57cec3cde488958e8bde2e1b764151","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a252ab3646661f1bb123fd760ae20127bd57cec3cde488958e8bde2e1b764151","first_computed_at":"2026-06-09T01:04:56.611171Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-09T01:04:56.611171Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"XwQ4UqF2UKEgMADGbdYlD8zk4IXNeNO3TxNnFHv6Zn8ttj9f5ErBNOO9JAMFv4ffpuvxxHXnWUFzT+ju856OAw==","signature_status":"signed_v1","signed_at":"2026-06-09T01:04:56.611597Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.07963","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a48b4240944b9b33825775f16e34b92a0c9d24c3536aa635a9bbf26dbae78e0a","sha256:6d035922c1d2e647d93455b6989208e6f4e4397d5672de1b79c4573a53625154"],"state_sha256":"82170e6939f8166d9b9512f803437c7dee164f92f985265938cc70149ddc04b6"}