{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:ZKI55P2ONZVIASEQIOJDYTDH7D","short_pith_number":"pith:ZKI55P2O","canonical_record":{"source":{"id":"2506.15740","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-06-17T15:46:15Z","cross_cats_sorted":["cs.CR","cs.LG"],"title_canon_sha256":"0ab624098d7cf0672a8207b72e3b32181f532e48e6f0bdfdb1b23339507bfc4a","abstract_canon_sha256":"c1d07cc2d9d82a3a6907c8ebab888e1c767c24319ce895aeacf0109c36e35bdf"},"schema_version":"1.0"},"canonical_sha256":"ca91debf4e6e6a80489043923c4c67f8cd419ff5f5d2cd5b352ecae843f06318","source":{"kind":"arxiv","id":"2506.15740","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.15740","created_at":"2026-07-05T11:34:09Z"},{"alias_kind":"arxiv_version","alias_value":"2506.15740v2","created_at":"2026-07-05T11:34:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.15740","created_at":"2026-07-05T11:34:09Z"},{"alias_kind":"pith_short_12","alias_value":"ZKI55P2ONZVI","created_at":"2026-07-05T11:34:09Z"},{"alias_kind":"pith_short_16","alias_value":"ZKI55P2ONZVIASEQ","created_at":"2026-07-05T11:34:09Z"},{"alias_kind":"pith_short_8","alias_value":"ZKI55P2O","created_at":"2026-07-05T11:34:09Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:ZKI55P2ONZVIASEQIOJDYTDH7D","target":"record","payload":{"canonical_record":{"source":{"id":"2506.15740","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-06-17T15:46:15Z","cross_cats_sorted":["cs.CR","cs.LG"],"title_canon_sha256":"0ab624098d7cf0672a8207b72e3b32181f532e48e6f0bdfdb1b23339507bfc4a","abstract_canon_sha256":"c1d07cc2d9d82a3a6907c8ebab888e1c767c24319ce895aeacf0109c36e35bdf"},"schema_version":"1.0"},"canonical_sha256":"ca91debf4e6e6a80489043923c4c67f8cd419ff5f5d2cd5b352ecae843f06318","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:34:09.578434Z","signature_b64":"MiAiBDQ08EyRbF7JCpIluIAC+X7w7n410803pXt7d/xFSbHM7rb4i1W02lrCugauU3js3YfpdO3yiujwx9NjDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ca91debf4e6e6a80489043923c4c67f8cd419ff5f5d2cd5b352ecae843f06318","last_reissued_at":"2026-07-05T11:34:09.577953Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:34:09.577953Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2506.15740","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:34:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ceYEV7ebAZ11d8vJ9DIldSE1ktQ2eRxdgAVqqE9UQVnDvEN/xoGqyQAby5CSUHch0xdfSeSFqj+pBaYFJupRDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T19:01:35.531891Z"},"content_sha256":"ec1fa3652b93085d455d32cbe4c6961d7ab0f2781fe5d835e0d4e6fcc12c374a","schema_version":"1.0","event_id":"sha256:ec1fa3652b93085d455d32cbe4c6961d7ab0f2781fe5d835e0d4e6fcc12c374a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:ZKI55P2ONZVIASEQIOJDYTDH7D","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"SHADE-Arena: Evaluating Sabotage and Monitoring in LLM Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR","cs.LG"],"primary_cat":"cs.AI","authors_text":"Buck Shlegeris, Chen Bo Calvin Zhang, Henry Sleight, Joe Benton, John Hughes, Jonathan Kutasov, Linda Petrini, Paul Colognese, Teun van der Weij, Tyler Tracy, Xiang Deng, Yuqi Sun","submitted_at":"2025-06-17T15:46:15Z","abstract_excerpt":"As Large Language Models (LLMs) are increasingly deployed as autonomous agents in complex and long horizon settings, it is critical to evaluate their ability to sabotage users by pursuing hidden objectives. We study the ability of frontier LLMs to evade monitoring and achieve harmful hidden goals while completing a wide array of realistic tasks. We evaluate a broad range of frontier LLMs using SHADE (Subtle Harmful Agent Detection & Evaluation)-Arena, the first highly diverse agent evaluation dataset for sabotage and monitoring capabilities of LLM agents. SHADE-Arena consists of complex pairs "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.15740","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/2506.15740/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":"2026-07-05T11:34:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"NjFPRWZksdIzMgju6A2JxGPtT9Kqo1yFWM08KD8teImeFgtkXP161kuZHddoYDhjnGFGzBC941VuOBBt2scGAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T19:01:35.532313Z"},"content_sha256":"123690d7eb2ee21134deea64baaa4a07faa320e2529616eba1ba3f96f2abc549","schema_version":"1.0","event_id":"sha256:123690d7eb2ee21134deea64baaa4a07faa320e2529616eba1ba3f96f2abc549"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/ZKI55P2ONZVIASEQIOJDYTDH7D/bundle.json","state_url":"https://pith.science/pith/ZKI55P2ONZVIASEQIOJDYTDH7D/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/ZKI55P2ONZVIASEQIOJDYTDH7D/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":"2026-07-06T19:01:35Z","links":{"resolver":"https://pith.science/pith/ZKI55P2ONZVIASEQIOJDYTDH7D","bundle":"https://pith.science/pith/ZKI55P2ONZVIASEQIOJDYTDH7D/bundle.json","state":"https://pith.science/pith/ZKI55P2ONZVIASEQIOJDYTDH7D/state.json","well_known_bundle":"https://pith.science/.well-known/pith/ZKI55P2ONZVIASEQIOJDYTDH7D/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:ZKI55P2ONZVIASEQIOJDYTDH7D","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":"c1d07cc2d9d82a3a6907c8ebab888e1c767c24319ce895aeacf0109c36e35bdf","cross_cats_sorted":["cs.CR","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-06-17T15:46:15Z","title_canon_sha256":"0ab624098d7cf0672a8207b72e3b32181f532e48e6f0bdfdb1b23339507bfc4a"},"schema_version":"1.0","source":{"id":"2506.15740","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.15740","created_at":"2026-07-05T11:34:09Z"},{"alias_kind":"arxiv_version","alias_value":"2506.15740v2","created_at":"2026-07-05T11:34:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.15740","created_at":"2026-07-05T11:34:09Z"},{"alias_kind":"pith_short_12","alias_value":"ZKI55P2ONZVI","created_at":"2026-07-05T11:34:09Z"},{"alias_kind":"pith_short_16","alias_value":"ZKI55P2ONZVIASEQ","created_at":"2026-07-05T11:34:09Z"},{"alias_kind":"pith_short_8","alias_value":"ZKI55P2O","created_at":"2026-07-05T11:34:09Z"}],"graph_snapshots":[{"event_id":"sha256:123690d7eb2ee21134deea64baaa4a07faa320e2529616eba1ba3f96f2abc549","target":"graph","created_at":"2026-07-05T11:34:09Z","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/2506.15740/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"As Large Language Models (LLMs) are increasingly deployed as autonomous agents in complex and long horizon settings, it is critical to evaluate their ability to sabotage users by pursuing hidden objectives. We study the ability of frontier LLMs to evade monitoring and achieve harmful hidden goals while completing a wide array of realistic tasks. We evaluate a broad range of frontier LLMs using SHADE (Subtle Harmful Agent Detection & Evaluation)-Arena, the first highly diverse agent evaluation dataset for sabotage and monitoring capabilities of LLM agents. SHADE-Arena consists of complex pairs ","authors_text":"Buck Shlegeris, Chen Bo Calvin Zhang, Henry Sleight, Joe Benton, John Hughes, Jonathan Kutasov, Linda Petrini, Paul Colognese, Teun van der Weij, Tyler Tracy, Xiang Deng, Yuqi Sun","cross_cats":["cs.CR","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-06-17T15:46:15Z","title":"SHADE-Arena: Evaluating Sabotage and Monitoring in LLM Agents"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.15740","kind":"arxiv","version":2},"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:ec1fa3652b93085d455d32cbe4c6961d7ab0f2781fe5d835e0d4e6fcc12c374a","target":"record","created_at":"2026-07-05T11:34:09Z","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":"c1d07cc2d9d82a3a6907c8ebab888e1c767c24319ce895aeacf0109c36e35bdf","cross_cats_sorted":["cs.CR","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-06-17T15:46:15Z","title_canon_sha256":"0ab624098d7cf0672a8207b72e3b32181f532e48e6f0bdfdb1b23339507bfc4a"},"schema_version":"1.0","source":{"id":"2506.15740","kind":"arxiv","version":2}},"canonical_sha256":"ca91debf4e6e6a80489043923c4c67f8cd419ff5f5d2cd5b352ecae843f06318","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ca91debf4e6e6a80489043923c4c67f8cd419ff5f5d2cd5b352ecae843f06318","first_computed_at":"2026-07-05T11:34:09.577953Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:34:09.577953Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"MiAiBDQ08EyRbF7JCpIluIAC+X7w7n410803pXt7d/xFSbHM7rb4i1W02lrCugauU3js3YfpdO3yiujwx9NjDA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:34:09.578434Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.15740","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:ec1fa3652b93085d455d32cbe4c6961d7ab0f2781fe5d835e0d4e6fcc12c374a","sha256:123690d7eb2ee21134deea64baaa4a07faa320e2529616eba1ba3f96f2abc549"],"state_sha256":"3ad66f9a869c13d3453ddcb7d31d48908812f2d168b137afbdc0d654a38dc833"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tfLT+RnxIvtqhu5HzT2oQR+BLK/WqL57MUVX0muB9XLFzQDBhcHokEE+Bq8rmOYyyUVrMhqOvlzAzXXnKZRRBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-06T19:01:35.534221Z","bundle_sha256":"1076daf2cca05331f3d33a2c9d556c70ed0c5550ca9cd8b9d4a682cd4098d837"}}