{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:Q3QWTFIPAKN5RPNEZHTR7SCXS6","short_pith_number":"pith:Q3QWTFIP","schema_version":"1.0","canonical_sha256":"86e169950f029bd8bda4c9e71fc85797beab9e3cd39f629ccbd51098c11ba246","source":{"kind":"arxiv","id":"2607.14285","version":1},"attestation_state":"computed","paper":{"title":"ToolAlignBench: Investigating Alignment Conflicts in Tool-Calling Enabled LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.SE","authors_text":"Amrita Bhattacharjee, Aryan Keluskar, Huan Liu","submitted_at":"2026-07-15T18:48:49Z","abstract_excerpt":"Safety alignment in LLMs aims to align models with human values, but which values take precedence when they conflict? We investigate this question in the context of tool-calling LLM agents deployed in regulated industries, where agents processing confidential documents may encounter content that triggers safety-trained values (e.g., public welfare) that conflict with deployment-context instructions (e.g., internal logging). To empirically verify this phenomenon, we build a benchmark of 128 scenarios across 16 domains. We find that safety-aligned open-source models override their deployment ins"},"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":"2607.14285","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2026-07-15T18:48:49Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7385ed195eedaedc34c4c0806bfbcc70e89a5efa57817a942a402dea6150b6ab","abstract_canon_sha256":"4731426017ef2f7458b7ef99887950b8e340a52704e16d33680a5cc2e165a584"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-17T00:21:03.278855Z","signature_b64":"M37lR+O4UwuhqatYhzL63T/X/m7au4LSOebSBARt8bq95ewIV8wNr0C3AUBfpgX/+gFJQ6mfa9/C2gIfu2K9Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"86e169950f029bd8bda4c9e71fc85797beab9e3cd39f629ccbd51098c11ba246","last_reissued_at":"2026-07-17T00:21:03.278048Z","signature_status":"signed_v1","first_computed_at":"2026-07-17T00:21:03.278048Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ToolAlignBench: Investigating Alignment Conflicts in Tool-Calling Enabled LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.SE","authors_text":"Amrita Bhattacharjee, Aryan Keluskar, Huan Liu","submitted_at":"2026-07-15T18:48:49Z","abstract_excerpt":"Safety alignment in LLMs aims to align models with human values, but which values take precedence when they conflict? We investigate this question in the context of tool-calling LLM agents deployed in regulated industries, where agents processing confidential documents may encounter content that triggers safety-trained values (e.g., public welfare) that conflict with deployment-context instructions (e.g., internal logging). To empirically verify this phenomenon, we build a benchmark of 128 scenarios across 16 domains. We find that safety-aligned open-source models override their deployment ins"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.14285","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/2607.14285/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":"2607.14285","created_at":"2026-07-17T00:21:03.278468+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.14285v1","created_at":"2026-07-17T00:21:03.278468+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.14285","created_at":"2026-07-17T00:21:03.278468+00:00"},{"alias_kind":"pith_short_12","alias_value":"Q3QWTFIPAKN5","created_at":"2026-07-17T00:21:03.278468+00:00"},{"alias_kind":"pith_short_16","alias_value":"Q3QWTFIPAKN5RPNE","created_at":"2026-07-17T00:21:03.278468+00:00"},{"alias_kind":"pith_short_8","alias_value":"Q3QWTFIP","created_at":"2026-07-17T00:21:03.278468+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/Q3QWTFIPAKN5RPNEZHTR7SCXS6","json":"https://pith.science/pith/Q3QWTFIPAKN5RPNEZHTR7SCXS6.json","graph_json":"https://pith.science/api/pith-number/Q3QWTFIPAKN5RPNEZHTR7SCXS6/graph.json","events_json":"https://pith.science/api/pith-number/Q3QWTFIPAKN5RPNEZHTR7SCXS6/events.json","paper":"https://pith.science/paper/Q3QWTFIP"},"agent_actions":{"view_html":"https://pith.science/pith/Q3QWTFIPAKN5RPNEZHTR7SCXS6","download_json":"https://pith.science/pith/Q3QWTFIPAKN5RPNEZHTR7SCXS6.json","view_paper":"https://pith.science/paper/Q3QWTFIP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.14285&json=true","fetch_graph":"https://pith.science/api/pith-number/Q3QWTFIPAKN5RPNEZHTR7SCXS6/graph.json","fetch_events":"https://pith.science/api/pith-number/Q3QWTFIPAKN5RPNEZHTR7SCXS6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Q3QWTFIPAKN5RPNEZHTR7SCXS6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Q3QWTFIPAKN5RPNEZHTR7SCXS6/action/storage_attestation","attest_author":"https://pith.science/pith/Q3QWTFIPAKN5RPNEZHTR7SCXS6/action/author_attestation","sign_citation":"https://pith.science/pith/Q3QWTFIPAKN5RPNEZHTR7SCXS6/action/citation_signature","submit_replication":"https://pith.science/pith/Q3QWTFIPAKN5RPNEZHTR7SCXS6/action/replication_record"}},"created_at":"2026-07-17T00:21:03.278468+00:00","updated_at":"2026-07-17T00:21:03.278468+00:00"}