{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:3MSJE7HOZDEPJ432EEGSBCSHOA","short_pith_number":"pith:3MSJE7HO","schema_version":"1.0","canonical_sha256":"db24927ceec8c8f4f37a210d208a477038cccb41685778c7e8decb79c41f9520","source":{"kind":"arxiv","id":"2506.14682","version":1},"attestation_state":"computed","paper":{"title":"AIRTBench: Measuring Autonomous AI Red Teaming Capabilities in Language Models","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Ads Dawson, Nick Landers, Rob Mulla, Shane Caldwell","submitted_at":"2025-06-17T16:19:06Z","abstract_excerpt":"We introduce AIRTBench, an AI red teaming benchmark for evaluating language models' ability to autonomously discover and exploit Artificial Intelligence and Machine Learning (AI/ML) security vulnerabilities. The benchmark consists of 70 realistic black-box capture-the-flag (CTF) challenges from the Crucible challenge environment on the Dreadnode platform, requiring models to write python code to interact with and compromise AI systems. Claude-3.7-Sonnet emerged as the clear leader, solving 43 challenges (61% of the total suite, 46.9% overall success rate), with Gemini-2.5-Pro following at 39 c"},"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":"2506.14682","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CR","submitted_at":"2025-06-17T16:19:06Z","cross_cats_sorted":[],"title_canon_sha256":"b19e11441be7883818081b5bac12cbf9477f223477f3bca48ffc6cb0e2a0eb2c","abstract_canon_sha256":"98c12305e88ebf9c501d17d743fe135d1391269217ad2a200a875ef845a71dc5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:23:00.725300Z","signature_b64":"6AsFkq5ukcAPyrsW2CSl4VZJ1GuEc9FPDyH85icozUqCsxzn1kyKeH8VVkMv1VRD4eauD4SzQXIe/DpIG1UFAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"db24927ceec8c8f4f37a210d208a477038cccb41685778c7e8decb79c41f9520","last_reissued_at":"2026-07-05T11:23:00.724774Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:23:00.724774Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AIRTBench: Measuring Autonomous AI Red Teaming Capabilities in Language Models","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Ads Dawson, Nick Landers, Rob Mulla, Shane Caldwell","submitted_at":"2025-06-17T16:19:06Z","abstract_excerpt":"We introduce AIRTBench, an AI red teaming benchmark for evaluating language models' ability to autonomously discover and exploit Artificial Intelligence and Machine Learning (AI/ML) security vulnerabilities. The benchmark consists of 70 realistic black-box capture-the-flag (CTF) challenges from the Crucible challenge environment on the Dreadnode platform, requiring models to write python code to interact with and compromise AI systems. Claude-3.7-Sonnet emerged as the clear leader, solving 43 challenges (61% of the total suite, 46.9% overall success rate), with Gemini-2.5-Pro following at 39 c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.14682","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/2506.14682/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":"2506.14682","created_at":"2026-07-05T11:23:00.724840+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.14682v1","created_at":"2026-07-05T11:23:00.724840+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.14682","created_at":"2026-07-05T11:23:00.724840+00:00"},{"alias_kind":"pith_short_12","alias_value":"3MSJE7HOZDEP","created_at":"2026-07-05T11:23:00.724840+00:00"},{"alias_kind":"pith_short_16","alias_value":"3MSJE7HOZDEPJ432","created_at":"2026-07-05T11:23:00.724840+00:00"},{"alias_kind":"pith_short_8","alias_value":"3MSJE7HO","created_at":"2026-07-05T11:23:00.724840+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.04019","citing_title":"Redefining AI Red Teaming in the Agentic Era: From Weeks to Hours","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3MSJE7HOZDEPJ432EEGSBCSHOA","json":"https://pith.science/pith/3MSJE7HOZDEPJ432EEGSBCSHOA.json","graph_json":"https://pith.science/api/pith-number/3MSJE7HOZDEPJ432EEGSBCSHOA/graph.json","events_json":"https://pith.science/api/pith-number/3MSJE7HOZDEPJ432EEGSBCSHOA/events.json","paper":"https://pith.science/paper/3MSJE7HO"},"agent_actions":{"view_html":"https://pith.science/pith/3MSJE7HOZDEPJ432EEGSBCSHOA","download_json":"https://pith.science/pith/3MSJE7HOZDEPJ432EEGSBCSHOA.json","view_paper":"https://pith.science/paper/3MSJE7HO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.14682&json=true","fetch_graph":"https://pith.science/api/pith-number/3MSJE7HOZDEPJ432EEGSBCSHOA/graph.json","fetch_events":"https://pith.science/api/pith-number/3MSJE7HOZDEPJ432EEGSBCSHOA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3MSJE7HOZDEPJ432EEGSBCSHOA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3MSJE7HOZDEPJ432EEGSBCSHOA/action/storage_attestation","attest_author":"https://pith.science/pith/3MSJE7HOZDEPJ432EEGSBCSHOA/action/author_attestation","sign_citation":"https://pith.science/pith/3MSJE7HOZDEPJ432EEGSBCSHOA/action/citation_signature","submit_replication":"https://pith.science/pith/3MSJE7HOZDEPJ432EEGSBCSHOA/action/replication_record"}},"created_at":"2026-07-05T11:23:00.724840+00:00","updated_at":"2026-07-05T11:23:00.724840+00:00"}