{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:53LZN4K4XPENLMADPYYQV4MPAT","short_pith_number":"pith:53LZN4K4","schema_version":"1.0","canonical_sha256":"eed796f15cbbc8d5b0037e310af18f04e0c51b11b9b0faeb9023e9acdefe1766","source":{"kind":"arxiv","id":"2506.02357","version":2},"attestation_state":"computed","paper":{"title":"Evaluating LLM Agent Adherence to Hierarchical Safety Principles: A Lightweight Benchmark for Probing Foundational Controllability Components","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CY"],"primary_cat":"cs.LG","authors_text":"Ram Potham (Independent Researcher)","submitted_at":"2025-06-03T01:16:34Z","abstract_excerpt":"Credible safety plans for advanced AI development require methods to verify agent behavior and detect potential control deficiencies early. A fundamental aspect is ensuring agents adhere to safety-critical principles, especially when these conflict with operational goals. This paper introduces a lightweight, interpretable benchmark to evaluate an LLM agent's ability to uphold a high-level safety principle when faced with conflicting task instructions. Our evaluation of six LLMs reveals two primary findings: (1) a quantifiable \"cost of compliance\" where safety constraints degrade task performan"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2506.02357","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-03T01:16:34Z","cross_cats_sorted":["cs.AI","cs.CY"],"title_canon_sha256":"abcf766c89488e63cd6a03eca755bddcd63dcfb301f1551605a7ee559f7b9bbe","abstract_canon_sha256":"9b9d95b162143d963387321c4fd59b404372db87204aa19bf2332d79b0db7f01"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:34:51.858933Z","signature_b64":"0ymUlYTW6pWldn+fedSvPwOVQV+HOskX+ygk8+ePRWNLzH9IEkwRgWZLwjPGr1TAhewoGevjf/BgeTLcnqs+Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eed796f15cbbc8d5b0037e310af18f04e0c51b11b9b0faeb9023e9acdefe1766","last_reissued_at":"2026-07-05T11:34:51.858428Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:34:51.858428Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Evaluating LLM Agent Adherence to Hierarchical Safety Principles: A Lightweight Benchmark for Probing Foundational Controllability Components","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CY"],"primary_cat":"cs.LG","authors_text":"Ram Potham (Independent Researcher)","submitted_at":"2025-06-03T01:16:34Z","abstract_excerpt":"Credible safety plans for advanced AI development require methods to verify agent behavior and detect potential control deficiencies early. A fundamental aspect is ensuring agents adhere to safety-critical principles, especially when these conflict with operational goals. This paper introduces a lightweight, interpretable benchmark to evaluate an LLM agent's ability to uphold a high-level safety principle when faced with conflicting task instructions. Our evaluation of six LLMs reveals two primary findings: (1) a quantifiable \"cost of compliance\" where safety constraints degrade task performan"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.02357","kind":"arxiv","version":2},"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/2506.02357/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2506.02357","created_at":"2026-07-05T11:34:51.858489+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.02357v2","created_at":"2026-07-05T11:34:51.858489+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.02357","created_at":"2026-07-05T11:34:51.858489+00:00"},{"alias_kind":"pith_short_12","alias_value":"53LZN4K4XPEN","created_at":"2026-07-05T11:34:51.858489+00:00"},{"alias_kind":"pith_short_16","alias_value":"53LZN4K4XPENLMAD","created_at":"2026-07-05T11:34:51.858489+00:00"},{"alias_kind":"pith_short_8","alias_value":"53LZN4K4","created_at":"2026-07-05T11:34:51.858489+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/53LZN4K4XPENLMADPYYQV4MPAT","json":"https://pith.science/pith/53LZN4K4XPENLMADPYYQV4MPAT.json","graph_json":"https://pith.science/api/pith-number/53LZN4K4XPENLMADPYYQV4MPAT/graph.json","events_json":"https://pith.science/api/pith-number/53LZN4K4XPENLMADPYYQV4MPAT/events.json","paper":"https://pith.science/paper/53LZN4K4"},"agent_actions":{"view_html":"https://pith.science/pith/53LZN4K4XPENLMADPYYQV4MPAT","download_json":"https://pith.science/pith/53LZN4K4XPENLMADPYYQV4MPAT.json","view_paper":"https://pith.science/paper/53LZN4K4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.02357&json=true","fetch_graph":"https://pith.science/api/pith-number/53LZN4K4XPENLMADPYYQV4MPAT/graph.json","fetch_events":"https://pith.science/api/pith-number/53LZN4K4XPENLMADPYYQV4MPAT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/53LZN4K4XPENLMADPYYQV4MPAT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/53LZN4K4XPENLMADPYYQV4MPAT/action/storage_attestation","attest_author":"https://pith.science/pith/53LZN4K4XPENLMADPYYQV4MPAT/action/author_attestation","sign_citation":"https://pith.science/pith/53LZN4K4XPENLMADPYYQV4MPAT/action/citation_signature","submit_replication":"https://pith.science/pith/53LZN4K4XPENLMADPYYQV4MPAT/action/replication_record"}},"created_at":"2026-07-05T11:34:51.858489+00:00","updated_at":"2026-07-05T11:34:51.858489+00:00"}