{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:ARDP5LAFFKUSZGXBNPCYJ5DMTT","short_pith_number":"pith:ARDP5LAF","canonical_record":{"source":{"id":"2506.04679","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-05T06:57:28Z","cross_cats_sorted":[],"title_canon_sha256":"1fac0a139d90a1ad1cb5ef9410c5ef21562fd91585a4c137609e8a262e9e3576","abstract_canon_sha256":"5c84da52fdd8c13cbce985f22a036b3f35486600feba4eb974d1a6197c06f08c"},"schema_version":"1.0"},"canonical_sha256":"0446feac052aa92c9ae16bc584f46c9cec4b8dba268d51e3d8e03f6f96f9fa35","source":{"kind":"arxiv","id":"2506.04679","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.04679","created_at":"2026-07-05T11:16:18Z"},{"alias_kind":"arxiv_version","alias_value":"2506.04679v1","created_at":"2026-07-05T11:16:18Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.04679","created_at":"2026-07-05T11:16:18Z"},{"alias_kind":"pith_short_12","alias_value":"ARDP5LAFFKUS","created_at":"2026-07-05T11:16:18Z"},{"alias_kind":"pith_short_16","alias_value":"ARDP5LAFFKUSZGXB","created_at":"2026-07-05T11:16:18Z"},{"alias_kind":"pith_short_8","alias_value":"ARDP5LAF","created_at":"2026-07-05T11:16:18Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:ARDP5LAFFKUSZGXBNPCYJ5DMTT","target":"record","payload":{"canonical_record":{"source":{"id":"2506.04679","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-05T06:57:28Z","cross_cats_sorted":[],"title_canon_sha256":"1fac0a139d90a1ad1cb5ef9410c5ef21562fd91585a4c137609e8a262e9e3576","abstract_canon_sha256":"5c84da52fdd8c13cbce985f22a036b3f35486600feba4eb974d1a6197c06f08c"},"schema_version":"1.0"},"canonical_sha256":"0446feac052aa92c9ae16bc584f46c9cec4b8dba268d51e3d8e03f6f96f9fa35","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:16:18.798417Z","signature_b64":"rTi9/3NSh/mVYtc6RnQfnRs4HLtt82A+vfxER/bZzaFCJkybuG9Tnzby3oKpkU+oQalvGwfBLWgtPe0yS4iEAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0446feac052aa92c9ae16bc584f46c9cec4b8dba268d51e3d8e03f6f96f9fa35","last_reissued_at":"2026-07-05T11:16:18.798006Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:16:18.798006Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2506.04679","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":"2026-07-05T11:16:18Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3jdiIVr5ARq0f1MsozjyBNQ1Srxo+iCELhCSvDpcB86sCasaIZF2uujM0hOnenqj9DJUmPgr72ShV+L5775WAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T02:54:23.837746Z"},"content_sha256":"3a7677787064d90bdd6f8d48d21012cf6f3f4c37f3d92815932fe3de381c5e29","schema_version":"1.0","event_id":"sha256:3a7677787064d90bdd6f8d48d21012cf6f3f4c37f3d92815932fe3de381c5e29"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:ARDP5LAFFKUSZGXBNPCYJ5DMTT","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Normative Conflicts and Shallow AI Alignment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Rapha\\\"el Milli\\`ere","submitted_at":"2025-06-05T06:57:28Z","abstract_excerpt":"The progress of AI systems such as large language models (LLMs) raises increasingly pressing concerns about their safe deployment. This paper examines the value alignment problem for LLMs, arguing that current alignment strategies are fundamentally inadequate to prevent misuse. Despite ongoing efforts to instill norms such as helpfulness, honesty, and harmlessness in LLMs through fine-tuning based on human preferences, they remain vulnerable to adversarial attacks that exploit conflicts between these norms. I argue that this vulnerability reflects a fundamental limitation of existing alignment"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.04679","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/2506.04679/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":"2026-07-05T11:16:18Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6z7l7dWTKv6r+ecuEs9+Eef6s+Zh+CY0c1QbjdDXQLh7fUO8NbS3AFxTu5xd0N024RNXTt/g2a5uK1iQ0hq0CQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T02:54:23.838240Z"},"content_sha256":"bf2a7487b3d9ab4a4f13dd5c5dd280637fb619e0bb7fb9b7c5eddd72f33b4eef","schema_version":"1.0","event_id":"sha256:bf2a7487b3d9ab4a4f13dd5c5dd280637fb619e0bb7fb9b7c5eddd72f33b4eef"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/ARDP5LAFFKUSZGXBNPCYJ5DMTT/bundle.json","state_url":"https://pith.science/pith/ARDP5LAFFKUSZGXBNPCYJ5DMTT/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/ARDP5LAFFKUSZGXBNPCYJ5DMTT/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":"2026-08-08T02:54:23Z","links":{"resolver":"https://pith.science/pith/ARDP5LAFFKUSZGXBNPCYJ5DMTT","bundle":"https://pith.science/pith/ARDP5LAFFKUSZGXBNPCYJ5DMTT/bundle.json","state":"https://pith.science/pith/ARDP5LAFFKUSZGXBNPCYJ5DMTT/state.json","well_known_bundle":"https://pith.science/.well-known/pith/ARDP5LAFFKUSZGXBNPCYJ5DMTT/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:ARDP5LAFFKUSZGXBNPCYJ5DMTT","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":"5c84da52fdd8c13cbce985f22a036b3f35486600feba4eb974d1a6197c06f08c","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-05T06:57:28Z","title_canon_sha256":"1fac0a139d90a1ad1cb5ef9410c5ef21562fd91585a4c137609e8a262e9e3576"},"schema_version":"1.0","source":{"id":"2506.04679","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.04679","created_at":"2026-07-05T11:16:18Z"},{"alias_kind":"arxiv_version","alias_value":"2506.04679v1","created_at":"2026-07-05T11:16:18Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.04679","created_at":"2026-07-05T11:16:18Z"},{"alias_kind":"pith_short_12","alias_value":"ARDP5LAFFKUS","created_at":"2026-07-05T11:16:18Z"},{"alias_kind":"pith_short_16","alias_value":"ARDP5LAFFKUSZGXB","created_at":"2026-07-05T11:16:18Z"},{"alias_kind":"pith_short_8","alias_value":"ARDP5LAF","created_at":"2026-07-05T11:16:18Z"}],"graph_snapshots":[{"event_id":"sha256:bf2a7487b3d9ab4a4f13dd5c5dd280637fb619e0bb7fb9b7c5eddd72f33b4eef","target":"graph","created_at":"2026-07-05T11:16:18Z","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/2506.04679/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The progress of AI systems such as large language models (LLMs) raises increasingly pressing concerns about their safe deployment. This paper examines the value alignment problem for LLMs, arguing that current alignment strategies are fundamentally inadequate to prevent misuse. Despite ongoing efforts to instill norms such as helpfulness, honesty, and harmlessness in LLMs through fine-tuning based on human preferences, they remain vulnerable to adversarial attacks that exploit conflicts between these norms. I argue that this vulnerability reflects a fundamental limitation of existing alignment","authors_text":"Rapha\\\"el Milli\\`ere","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-05T06:57:28Z","title":"Normative Conflicts and Shallow AI Alignment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.04679","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":"sha256:3a7677787064d90bdd6f8d48d21012cf6f3f4c37f3d92815932fe3de381c5e29","target":"record","created_at":"2026-07-05T11:16:18Z","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":"5c84da52fdd8c13cbce985f22a036b3f35486600feba4eb974d1a6197c06f08c","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-05T06:57:28Z","title_canon_sha256":"1fac0a139d90a1ad1cb5ef9410c5ef21562fd91585a4c137609e8a262e9e3576"},"schema_version":"1.0","source":{"id":"2506.04679","kind":"arxiv","version":1}},"canonical_sha256":"0446feac052aa92c9ae16bc584f46c9cec4b8dba268d51e3d8e03f6f96f9fa35","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0446feac052aa92c9ae16bc584f46c9cec4b8dba268d51e3d8e03f6f96f9fa35","first_computed_at":"2026-07-05T11:16:18.798006Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:16:18.798006Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"rTi9/3NSh/mVYtc6RnQfnRs4HLtt82A+vfxER/bZzaFCJkybuG9Tnzby3oKpkU+oQalvGwfBLWgtPe0yS4iEAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:16:18.798417Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.04679","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3a7677787064d90bdd6f8d48d21012cf6f3f4c37f3d92815932fe3de381c5e29","sha256:bf2a7487b3d9ab4a4f13dd5c5dd280637fb619e0bb7fb9b7c5eddd72f33b4eef"],"state_sha256":"9532d0d89c4227b5188aab91adc0436a3c6e4abb819f618b7c084a945dd2e29e"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"DPy1B/v80IcNvPocVhXfeo+JZgF4kUV31vVWkPbNOrMsL9yKaLKSIUMygQfJO8/xLfBuEGcbsQkolK0ZGAJEAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T02:54:23.841564Z","bundle_sha256":"5b62d62d3d699ed759627567733c799d58e6dfe94779cd69c96f1da64b7d646c"}}