{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:NFOPMADRNNNPKRR7UFENCNEXJY","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":"304e20d165bad41856f239bee7f6cbb50dcf546514e6b078c5f0409692056b12","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-13T04:08:49Z","title_canon_sha256":"fbb3b845fe71aee1af8e8f733ce98957fae3494ec3d48b5bcd288f3bfd76fc97"},"schema_version":"1.0","source":{"id":"2311.07064","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2311.07064","created_at":"2026-07-05T09:16:10Z"},{"alias_kind":"arxiv_version","alias_value":"2311.07064v3","created_at":"2026-07-05T09:16:10Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.07064","created_at":"2026-07-05T09:16:10Z"},{"alias_kind":"pith_short_12","alias_value":"NFOPMADRNNNP","created_at":"2026-07-05T09:16:10Z"},{"alias_kind":"pith_short_16","alias_value":"NFOPMADRNNNPKRR7","created_at":"2026-07-05T09:16:10Z"},{"alias_kind":"pith_short_8","alias_value":"NFOPMADR","created_at":"2026-07-05T09:16:10Z"}],"graph_snapshots":[{"event_id":"sha256:1c2bb092284526070d4bd009c8b20f70cb951db977ddadb8dcda440699fdee57","target":"graph","created_at":"2026-07-05T09:16:10Z","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/2311.07064/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We discover that many natural-language prompts can be replaced by corresponding prompts that are unintelligible to humans but that provably elicit similar behavior in language models. We call these prompts \"evil twins\" because they are obfuscated and uninterpretable (evil), but at the same time mimic the functionality of the original natural-language prompts (twins). Remarkably, evil twins transfer between models. We find these prompts by solving a maximum-likelihood problem which has applications of independent interest.","authors_text":"Enric Boix-Adsera, H. Howie Huang, Lucas H. McCabe, Rimon Melamed, Tanay Wakhare, Yejin Kim","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-13T04:08:49Z","title":"Prompts have evil twins"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.07064","kind":"arxiv","version":3},"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":"sha256:484c92457e6055878118ae05fc47b62c55469abc6f651dd96ab6497553b244be","target":"record","created_at":"2026-07-05T09:16:10Z","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":"304e20d165bad41856f239bee7f6cbb50dcf546514e6b078c5f0409692056b12","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-13T04:08:49Z","title_canon_sha256":"fbb3b845fe71aee1af8e8f733ce98957fae3494ec3d48b5bcd288f3bfd76fc97"},"schema_version":"1.0","source":{"id":"2311.07064","kind":"arxiv","version":3}},"canonical_sha256":"695cf600716b5af5463fa148d134974e3a0c3cfddb05af0e453ac458851b3666","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"695cf600716b5af5463fa148d134974e3a0c3cfddb05af0e453ac458851b3666","first_computed_at":"2026-07-05T09:16:10.072619Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:16:10.072619Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"BLc8F/wM3CUFolQ9Tjm1CxI07TXEq/Z+eiU+dQsRCIbsQKmJymgsyPYzi+cJA+K4285v785Dz/q8XJRzOI4yAA==","signature_status":"signed_v1","signed_at":"2026-07-05T09:16:10.073060Z","signed_message":"canonical_sha256_bytes"},"source_id":"2311.07064","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:484c92457e6055878118ae05fc47b62c55469abc6f651dd96ab6497553b244be","sha256:1c2bb092284526070d4bd009c8b20f70cb951db977ddadb8dcda440699fdee57"],"state_sha256":"35ddbd44bef899d443a6325ba6fca95f6ee3ed5073bcd52e9940b3b6773eb700"}