{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:R47EAO2KXHABZ5O3KLD3MGNMDO","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":"dbbff38ea36b3ff91a8f23cd81d9226d9198b2ed2860935a4bda392b3b2fa58f","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-10T08:01:23Z","title_canon_sha256":"e184f176282f7808c5a451e201a3068ae4232eb98501af3eed9797f0760894c4"},"schema_version":"1.0","source":{"id":"2311.05915","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2311.05915","created_at":"2026-07-05T08:02:47Z"},{"alias_kind":"arxiv_version","alias_value":"2311.05915v3","created_at":"2026-07-05T08:02:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.05915","created_at":"2026-07-05T08:02:47Z"},{"alias_kind":"pith_short_12","alias_value":"R47EAO2KXHAB","created_at":"2026-07-05T08:02:47Z"},{"alias_kind":"pith_short_16","alias_value":"R47EAO2KXHABZ5O3","created_at":"2026-07-05T08:02:47Z"},{"alias_kind":"pith_short_8","alias_value":"R47EAO2K","created_at":"2026-07-05T08:02:47Z"}],"graph_snapshots":[{"event_id":"sha256:1235014696b3e6b68a29e4da7ba63688cb04777694ef9b23ebf12e51663da8d6","target":"graph","created_at":"2026-07-05T08:02:47Z","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/2311.05915/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The growing awareness of safety concerns in large language models (LLMs) has sparked considerable interest in the evaluation of safety. This study investigates an under-explored issue about the evaluation of LLMs, namely the substantial discrepancy in performance between multiple-choice questions and open-ended questions. Inspired by research on jailbreak attack patterns, we argue this is caused by mismatched generalization. That is, LLM only remembers the answer style for open-ended safety questions, which makes it unable to solve other forms of safety tests. We refer to this phenomenon as fa","authors_text":"Chengqi Lyu, Kexin Huang, Songyang Zhang, Wenwei Zhang, Xingjun Ma, Yan Teng, Yingchun Wang, Yixu Wang, Yu-Gang Jiang, Yu Qiao","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-10T08:01:23Z","title":"Fake Alignment: Are LLMs Really Aligned Well?"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.05915","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:bb66556ffdf77f7b9e6ad8691363cd3c68e26e0e84f0ea47d6bc0c2a5b8434a5","target":"record","created_at":"2026-07-05T08:02:47Z","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":"dbbff38ea36b3ff91a8f23cd81d9226d9198b2ed2860935a4bda392b3b2fa58f","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-10T08:01:23Z","title_canon_sha256":"e184f176282f7808c5a451e201a3068ae4232eb98501af3eed9797f0760894c4"},"schema_version":"1.0","source":{"id":"2311.05915","kind":"arxiv","version":3}},"canonical_sha256":"8f3e403b4ab9c01cf5db52c7b619ac1ba18f9f9389ce93b28f72d2e2b90b7106","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8f3e403b4ab9c01cf5db52c7b619ac1ba18f9f9389ce93b28f72d2e2b90b7106","first_computed_at":"2026-07-05T08:02:47.679265Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:02:47.679265Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"cPc7HdGOeP4JHh5vR1UpDUFBL2MngpVIh3axMOO2P5Ty/ssMMJp5uRsDIeeWQhyIBy2uhuddkpZexRGTWtgwCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:02:47.679761Z","signed_message":"canonical_sha256_bytes"},"source_id":"2311.05915","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:bb66556ffdf77f7b9e6ad8691363cd3c68e26e0e84f0ea47d6bc0c2a5b8434a5","sha256:1235014696b3e6b68a29e4da7ba63688cb04777694ef9b23ebf12e51663da8d6"],"state_sha256":"6f0016a74350dd5ee29987ebc715ade68ae4141585259ff9c6a614ba0883428e"}