{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5HCL7PVZJQX7KAZYP657WEO3MW","short_pith_number":"pith:5HCL7PVZ","schema_version":"1.0","canonical_sha256":"e9c4bfbeb94c2ff503387fbbfb11db659bd554df5e129c38c8a61fb05d75e4eb","source":{"kind":"arxiv","id":"2507.06282","version":1},"attestation_state":"computed","paper":{"title":"The bitter lesson of misuse detection","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Charbel-Rapha\\\"el Segerie, Diego Dorn, Hadrien Mariaccia","submitted_at":"2025-07-08T15:21:17Z","abstract_excerpt":"Prior work on jailbreak detection has established the importance of adversarial robustness for LLMs but has largely focused on the model ability to resist adversarial inputs and to output safe content, rather than the effectiveness of external supervision systems. The only public and independent benchmark of these guardrails to date evaluates a narrow set of supervisors on limited scenarios. Consequently, no comprehensive public benchmark yet verifies how well supervision systems from the market perform under realistic, diverse attacks. To address this, we introduce BELLS, a Benchmark for the "},"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":"2507.06282","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CR","submitted_at":"2025-07-08T15:21:17Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"147026ee0e6147e09d26c255772966761d411d28bbadc7102b9f2c8691c72fce","abstract_canon_sha256":"25760b207398cac48954873660fed63a7bc24258f972d310245da45bc6a2391e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:34:13.019913Z","signature_b64":"tvSc3fEL0KLgd+9DYxa5iuusMZhdKX8hjoTpPetcTOaniPeFpjC8ymGX5QpBCDhkL4Tq5LJ8BS+U1vcUz1uYBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e9c4bfbeb94c2ff503387fbbfb11db659bd554df5e129c38c8a61fb05d75e4eb","last_reissued_at":"2026-07-05T11:34:13.019473Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:34:13.019473Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The bitter lesson of misuse detection","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Charbel-Rapha\\\"el Segerie, Diego Dorn, Hadrien Mariaccia","submitted_at":"2025-07-08T15:21:17Z","abstract_excerpt":"Prior work on jailbreak detection has established the importance of adversarial robustness for LLMs but has largely focused on the model ability to resist adversarial inputs and to output safe content, rather than the effectiveness of external supervision systems. The only public and independent benchmark of these guardrails to date evaluates a narrow set of supervisors on limited scenarios. Consequently, no comprehensive public benchmark yet verifies how well supervision systems from the market perform under realistic, diverse attacks. To address this, we introduce BELLS, a Benchmark for the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.06282","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/2507.06282/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":"2507.06282","created_at":"2026-07-05T11:34:13.019529+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.06282v1","created_at":"2026-07-05T11:34:13.019529+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.06282","created_at":"2026-07-05T11:34:13.019529+00:00"},{"alias_kind":"pith_short_12","alias_value":"5HCL7PVZJQX7","created_at":"2026-07-05T11:34:13.019529+00:00"},{"alias_kind":"pith_short_16","alias_value":"5HCL7PVZJQX7KAZY","created_at":"2026-07-05T11:34:13.019529+00:00"},{"alias_kind":"pith_short_8","alias_value":"5HCL7PVZ","created_at":"2026-07-05T11:34:13.019529+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/5HCL7PVZJQX7KAZYP657WEO3MW","json":"https://pith.science/pith/5HCL7PVZJQX7KAZYP657WEO3MW.json","graph_json":"https://pith.science/api/pith-number/5HCL7PVZJQX7KAZYP657WEO3MW/graph.json","events_json":"https://pith.science/api/pith-number/5HCL7PVZJQX7KAZYP657WEO3MW/events.json","paper":"https://pith.science/paper/5HCL7PVZ"},"agent_actions":{"view_html":"https://pith.science/pith/5HCL7PVZJQX7KAZYP657WEO3MW","download_json":"https://pith.science/pith/5HCL7PVZJQX7KAZYP657WEO3MW.json","view_paper":"https://pith.science/paper/5HCL7PVZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.06282&json=true","fetch_graph":"https://pith.science/api/pith-number/5HCL7PVZJQX7KAZYP657WEO3MW/graph.json","fetch_events":"https://pith.science/api/pith-number/5HCL7PVZJQX7KAZYP657WEO3MW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5HCL7PVZJQX7KAZYP657WEO3MW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5HCL7PVZJQX7KAZYP657WEO3MW/action/storage_attestation","attest_author":"https://pith.science/pith/5HCL7PVZJQX7KAZYP657WEO3MW/action/author_attestation","sign_citation":"https://pith.science/pith/5HCL7PVZJQX7KAZYP657WEO3MW/action/citation_signature","submit_replication":"https://pith.science/pith/5HCL7PVZJQX7KAZYP657WEO3MW/action/replication_record"}},"created_at":"2026-07-05T11:34:13.019529+00:00","updated_at":"2026-07-05T11:34:13.019529+00:00"}