{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:GXKWPWVTZGNXIB3IBBY7DG76AG","short_pith_number":"pith:GXKWPWVT","schema_version":"1.0","canonical_sha256":"35d567dab3c99b7407680871f19bfe01a0b9d9ee02e2e7f9ca8b1cb1554114dd","source":{"kind":"arxiv","id":"2606.25899","version":1},"attestation_state":"computed","paper":{"title":"Manipulation Is Task-Dependent: A Multi-Axis, Multi-Environment Evaluation of Frontier LLMs","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.MA","authors_text":"Adeeb Zaman, Erik Nordby, Fred Heiding","submitted_at":"2026-06-24T14:47:47Z","abstract_excerpt":"We evaluate manipulative behavior in six frontier language models across six environments, ranging from negotiation tasks to agentic workflows, resulting in 13{,}590 individual scenarios. Manipulation rates are measured across three axes: framing (mandate honesty or permit manipulation), incentive structure (from no incentives to substantial ones), and task difficulty. Existing benchmarks typically vary a single axis within a single environment, an approach our results show is insufficient. We rank models by manipulation rate and find Spearman rank correlations across environments average $\\rh"},"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":"2606.25899","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.MA","submitted_at":"2026-06-24T14:47:47Z","cross_cats_sorted":[],"title_canon_sha256":"babada62512dd42cbf92649edaf8cfca7640407f1f9276133e79039c0f942c1e","abstract_canon_sha256":"d4d8e21f7a408c7489330cbeb851998df35403dcc78998f3e3aca4be8eee4f83"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-25T01:18:42.674648Z","signature_b64":"2movlxw0GGUaGLs6yN23FOabrBM1OXdtFXYCjgumn9zsDoWS18YKIDazkMXJpzDLbmH0cQljYto7lDjCvxeMBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"35d567dab3c99b7407680871f19bfe01a0b9d9ee02e2e7f9ca8b1cb1554114dd","last_reissued_at":"2026-06-25T01:18:42.674154Z","signature_status":"signed_v1","first_computed_at":"2026-06-25T01:18:42.674154Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Manipulation Is Task-Dependent: A Multi-Axis, Multi-Environment Evaluation of Frontier LLMs","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.MA","authors_text":"Adeeb Zaman, Erik Nordby, Fred Heiding","submitted_at":"2026-06-24T14:47:47Z","abstract_excerpt":"We evaluate manipulative behavior in six frontier language models across six environments, ranging from negotiation tasks to agentic workflows, resulting in 13{,}590 individual scenarios. Manipulation rates are measured across three axes: framing (mandate honesty or permit manipulation), incentive structure (from no incentives to substantial ones), and task difficulty. Existing benchmarks typically vary a single axis within a single environment, an approach our results show is insufficient. We rank models by manipulation rate and find Spearman rank correlations across environments average $\\rh"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.25899","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/2606.25899/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":"2606.25899","created_at":"2026-06-25T01:18:42.674212+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.25899v1","created_at":"2026-06-25T01:18:42.674212+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.25899","created_at":"2026-06-25T01:18:42.674212+00:00"},{"alias_kind":"pith_short_12","alias_value":"GXKWPWVTZGNX","created_at":"2026-06-25T01:18:42.674212+00:00"},{"alias_kind":"pith_short_16","alias_value":"GXKWPWVTZGNXIB3I","created_at":"2026-06-25T01:18:42.674212+00:00"},{"alias_kind":"pith_short_8","alias_value":"GXKWPWVT","created_at":"2026-06-25T01:18:42.674212+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/GXKWPWVTZGNXIB3IBBY7DG76AG","json":"https://pith.science/pith/GXKWPWVTZGNXIB3IBBY7DG76AG.json","graph_json":"https://pith.science/api/pith-number/GXKWPWVTZGNXIB3IBBY7DG76AG/graph.json","events_json":"https://pith.science/api/pith-number/GXKWPWVTZGNXIB3IBBY7DG76AG/events.json","paper":"https://pith.science/paper/GXKWPWVT"},"agent_actions":{"view_html":"https://pith.science/pith/GXKWPWVTZGNXIB3IBBY7DG76AG","download_json":"https://pith.science/pith/GXKWPWVTZGNXIB3IBBY7DG76AG.json","view_paper":"https://pith.science/paper/GXKWPWVT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.25899&json=true","fetch_graph":"https://pith.science/api/pith-number/GXKWPWVTZGNXIB3IBBY7DG76AG/graph.json","fetch_events":"https://pith.science/api/pith-number/GXKWPWVTZGNXIB3IBBY7DG76AG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GXKWPWVTZGNXIB3IBBY7DG76AG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GXKWPWVTZGNXIB3IBBY7DG76AG/action/storage_attestation","attest_author":"https://pith.science/pith/GXKWPWVTZGNXIB3IBBY7DG76AG/action/author_attestation","sign_citation":"https://pith.science/pith/GXKWPWVTZGNXIB3IBBY7DG76AG/action/citation_signature","submit_replication":"https://pith.science/pith/GXKWPWVTZGNXIB3IBBY7DG76AG/action/replication_record"}},"created_at":"2026-06-25T01:18:42.674212+00:00","updated_at":"2026-06-25T01:18:42.674212+00:00"}