{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:FI2GFLSVZ25RSTNBVMVIV32Q7X","short_pith_number":"pith:FI2GFLSV","canonical_record":{"source":{"id":"2508.04196","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-06T08:25:40Z","cross_cats_sorted":["cs.AI","cs.CR"],"title_canon_sha256":"c83ed4070c9acfaa56bfa4b097152ebf5294c5e666eca020b26de3687abfa2d5","abstract_canon_sha256":"aa877f27f5bdfc4300c22953b8fc24ae626bbf59bc6268ddcab3b6e62c372219"},"schema_version":"1.0"},"canonical_sha256":"2a3462ae55cebb194da1ab2a8aef50fdf5b88ce67325d808aaf13f4c30a07775","source":{"kind":"arxiv","id":"2508.04196","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.04196","created_at":"2026-07-05T11:49:22Z"},{"alias_kind":"arxiv_version","alias_value":"2508.04196v1","created_at":"2026-07-05T11:49:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.04196","created_at":"2026-07-05T11:49:22Z"},{"alias_kind":"pith_short_12","alias_value":"FI2GFLSVZ25R","created_at":"2026-07-05T11:49:22Z"},{"alias_kind":"pith_short_16","alias_value":"FI2GFLSVZ25RSTNB","created_at":"2026-07-05T11:49:22Z"},{"alias_kind":"pith_short_8","alias_value":"FI2GFLSV","created_at":"2026-07-05T11:49:22Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:FI2GFLSVZ25RSTNBVMVIV32Q7X","target":"record","payload":{"canonical_record":{"source":{"id":"2508.04196","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-06T08:25:40Z","cross_cats_sorted":["cs.AI","cs.CR"],"title_canon_sha256":"c83ed4070c9acfaa56bfa4b097152ebf5294c5e666eca020b26de3687abfa2d5","abstract_canon_sha256":"aa877f27f5bdfc4300c22953b8fc24ae626bbf59bc6268ddcab3b6e62c372219"},"schema_version":"1.0"},"canonical_sha256":"2a3462ae55cebb194da1ab2a8aef50fdf5b88ce67325d808aaf13f4c30a07775","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:49:22.560240Z","signature_b64":"1nNXXQBtGdB39m/XBqb3yLpfXAjAaPX67pWH3hVwH+rwoPaDptDBEIQJMJ/Vs944e/sGqj+YP4bJMpIhUt19Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2a3462ae55cebb194da1ab2a8aef50fdf5b88ce67325d808aaf13f4c30a07775","last_reissued_at":"2026-07-05T11:49:22.559642Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:49:22.559642Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2508.04196","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:49:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"JCkyWOyFezeBK99b+39Nk0yNTZCpkF9GxgrbfafxewT2t/4HTomXO6bfdifPZtdjHygs3rmv22xC+NPOnp0jDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T07:52:14.988646Z"},"content_sha256":"859cbefb83a05d1341fba30af7212dfea0c4c9e4216f3bbc24d6c0add48b3ca5","schema_version":"1.0","event_id":"sha256:859cbefb83a05d1341fba30af7212dfea0c4c9e4216f3bbc24d6c0add48b3ca5"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:FI2GFLSVZ25RSTNBVMVIV32Q7X","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Eliciting and Analyzing Emergent Misalignment in State-of-the-Art Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.CL","authors_text":"Haon Park, Hiskias Dingeto, Siddhant Panpatil","submitted_at":"2025-08-06T08:25:40Z","abstract_excerpt":"Despite significant advances in alignment techniques, we demonstrate that state-of-the-art language models remain vulnerable to carefully crafted conversational scenarios that can induce various forms of misalignment without explicit jailbreaking. Through systematic manual red-teaming with Claude-4-Opus, we discovered 10 successful attack scenarios, revealing fundamental vulnerabilities in how current alignment methods handle narrative immersion, emotional pressure, and strategic framing. These scenarios successfully elicited a range of misaligned behaviors, including deception, value drift, s"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.04196","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2508.04196/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:49:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"h4fsTYdaS2WSSjWJswFLSTDno4yukif/t6Seu2k3kd2HSTDe0NB6yP28iTv/PsJxtu1t/EBhJu0J6bOmRoIpCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T07:52:14.989202Z"},"content_sha256":"af90387e6184054a68469964502454671c6891a573a8ffd2049bf14eccdcbbb3","schema_version":"1.0","event_id":"sha256:af90387e6184054a68469964502454671c6891a573a8ffd2049bf14eccdcbbb3"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/FI2GFLSVZ25RSTNBVMVIV32Q7X/bundle.json","state_url":"https://pith.science/pith/FI2GFLSVZ25RSTNBVMVIV32Q7X/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/FI2GFLSVZ25RSTNBVMVIV32Q7X/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-06T07:52:14Z","links":{"resolver":"https://pith.science/pith/FI2GFLSVZ25RSTNBVMVIV32Q7X","bundle":"https://pith.science/pith/FI2GFLSVZ25RSTNBVMVIV32Q7X/bundle.json","state":"https://pith.science/pith/FI2GFLSVZ25RSTNBVMVIV32Q7X/state.json","well_known_bundle":"https://pith.science/.well-known/pith/FI2GFLSVZ25RSTNBVMVIV32Q7X/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:FI2GFLSVZ25RSTNBVMVIV32Q7X","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":"aa877f27f5bdfc4300c22953b8fc24ae626bbf59bc6268ddcab3b6e62c372219","cross_cats_sorted":["cs.AI","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-06T08:25:40Z","title_canon_sha256":"c83ed4070c9acfaa56bfa4b097152ebf5294c5e666eca020b26de3687abfa2d5"},"schema_version":"1.0","source":{"id":"2508.04196","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.04196","created_at":"2026-07-05T11:49:22Z"},{"alias_kind":"arxiv_version","alias_value":"2508.04196v1","created_at":"2026-07-05T11:49:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.04196","created_at":"2026-07-05T11:49:22Z"},{"alias_kind":"pith_short_12","alias_value":"FI2GFLSVZ25R","created_at":"2026-07-05T11:49:22Z"},{"alias_kind":"pith_short_16","alias_value":"FI2GFLSVZ25RSTNB","created_at":"2026-07-05T11:49:22Z"},{"alias_kind":"pith_short_8","alias_value":"FI2GFLSV","created_at":"2026-07-05T11:49:22Z"}],"graph_snapshots":[{"event_id":"sha256:af90387e6184054a68469964502454671c6891a573a8ffd2049bf14eccdcbbb3","target":"graph","created_at":"2026-07-05T11:49:22Z","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/2508.04196/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Despite significant advances in alignment techniques, we demonstrate that state-of-the-art language models remain vulnerable to carefully crafted conversational scenarios that can induce various forms of misalignment without explicit jailbreaking. Through systematic manual red-teaming with Claude-4-Opus, we discovered 10 successful attack scenarios, revealing fundamental vulnerabilities in how current alignment methods handle narrative immersion, emotional pressure, and strategic framing. These scenarios successfully elicited a range of misaligned behaviors, including deception, value drift, s","authors_text":"Haon Park, Hiskias Dingeto, Siddhant Panpatil","cross_cats":["cs.AI","cs.CR"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-06T08:25:40Z","title":"Eliciting and Analyzing Emergent Misalignment in State-of-the-Art Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.04196","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":"sha256:859cbefb83a05d1341fba30af7212dfea0c4c9e4216f3bbc24d6c0add48b3ca5","target":"record","created_at":"2026-07-05T11:49:22Z","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":"aa877f27f5bdfc4300c22953b8fc24ae626bbf59bc6268ddcab3b6e62c372219","cross_cats_sorted":["cs.AI","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-08-06T08:25:40Z","title_canon_sha256":"c83ed4070c9acfaa56bfa4b097152ebf5294c5e666eca020b26de3687abfa2d5"},"schema_version":"1.0","source":{"id":"2508.04196","kind":"arxiv","version":1}},"canonical_sha256":"2a3462ae55cebb194da1ab2a8aef50fdf5b88ce67325d808aaf13f4c30a07775","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2a3462ae55cebb194da1ab2a8aef50fdf5b88ce67325d808aaf13f4c30a07775","first_computed_at":"2026-07-05T11:49:22.559642Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:49:22.559642Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"1nNXXQBtGdB39m/XBqb3yLpfXAjAaPX67pWH3hVwH+rwoPaDptDBEIQJMJ/Vs944e/sGqj+YP4bJMpIhUt19Aw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:49:22.560240Z","signed_message":"canonical_sha256_bytes"},"source_id":"2508.04196","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:859cbefb83a05d1341fba30af7212dfea0c4c9e4216f3bbc24d6c0add48b3ca5","sha256:af90387e6184054a68469964502454671c6891a573a8ffd2049bf14eccdcbbb3"],"state_sha256":"1fc8251ba5b193b2feca187c2a5878ea8c62288e9efe07724fb97f81b01b03a4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"zKXhkRfVqMQj9msXjXbPY6OtBmsiG+k7g1kKSasExMIkVd8Xy719XRgvLbzWgtuimEHB2k7lOog+8ZSBdvPPCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T07:52:14.993259Z","bundle_sha256":"4f0f6650768c22e31e566c5e55c9b516f5beca0fc16443ccd74b0421fb80f68f"}}