{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:OF5DIUW5P3YBYY6NOHLSER5LXN","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":"bee9124b69989429e47fa51f97d1b4b8ded45096b68e74ef652e85ab45027ad5","cross_cats_sorted":["cs.AI","cs.CL","cs.HC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-06T11:57:42Z","title_canon_sha256":"5f3681b83f3767da5c6624df2634422d6cfa250d9c35ff08e338a3dc40782bdd"},"schema_version":"1.0","source":{"id":"2507.06253","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.06253","created_at":"2026-07-05T11:34:12Z"},{"alias_kind":"arxiv_version","alias_value":"2507.06253v1","created_at":"2026-07-05T11:34:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.06253","created_at":"2026-07-05T11:34:12Z"},{"alias_kind":"pith_short_12","alias_value":"OF5DIUW5P3YB","created_at":"2026-07-05T11:34:12Z"},{"alias_kind":"pith_short_16","alias_value":"OF5DIUW5P3YBYY6N","created_at":"2026-07-05T11:34:12Z"},{"alias_kind":"pith_short_8","alias_value":"OF5DIUW5","created_at":"2026-07-05T11:34:12Z"}],"graph_snapshots":[{"event_id":"sha256:f52e17afdc51c01f0d11b72d17a7bf01e5993c74f587940dc2b829fcb5345e13","target":"graph","created_at":"2026-07-05T11:34:12Z","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/2507.06253/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Betley et al. (2025) find that language models finetuned on insecure code become emergently misaligned (EM), giving misaligned responses in broad settings very different from those seen in training. However, it remains unclear as to why emergent misalignment occurs.\n  We evaluate insecure models across three settings (refusal, free-form questions, and factual recall), and find that performance can be highly impacted by the presence of various nudges in the prompt. In the refusal and free-form questions, we find that we can reliably elicit misaligned behaviour from insecure models simply by ask","authors_text":"Anna Soligo, Daniel Tan, Tim Wyse, Twm Stone","cross_cats":["cs.AI","cs.CL","cs.HC"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-06T11:57:42Z","title":"Emergent misalignment as prompt sensitivity: A research note"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.06253","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:b900138001b849ecb0ab1864c7d5b86a57783540a91f27af4e48a7513ab777f6","target":"record","created_at":"2026-07-05T11:34:12Z","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":"bee9124b69989429e47fa51f97d1b4b8ded45096b68e74ef652e85ab45027ad5","cross_cats_sorted":["cs.AI","cs.CL","cs.HC"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-06T11:57:42Z","title_canon_sha256":"5f3681b83f3767da5c6624df2634422d6cfa250d9c35ff08e338a3dc40782bdd"},"schema_version":"1.0","source":{"id":"2507.06253","kind":"arxiv","version":1}},"canonical_sha256":"717a3452dd7ef01c63cd71d72247abbb54b659545ffb4f537583dd011ccc5b81","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"717a3452dd7ef01c63cd71d72247abbb54b659545ffb4f537583dd011ccc5b81","first_computed_at":"2026-07-05T11:34:12.596522Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:34:12.596522Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"HmgdcDEHCZNo51KsBjwkkxUze/rAMa0KgqwJhNtr+9xW3uZBNbpJkNuF2Mcfq9DlvBh3ky/fhbmmcdvOZknDDA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:34:12.597010Z","signed_message":"canonical_sha256_bytes"},"source_id":"2507.06253","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b900138001b849ecb0ab1864c7d5b86a57783540a91f27af4e48a7513ab777f6","sha256:f52e17afdc51c01f0d11b72d17a7bf01e5993c74f587940dc2b829fcb5345e13"],"state_sha256":"17daca47d093c474c73ae507bf4b93fb36bc3cb2cbc2e0f80519406351caf13e"}