{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:7MY4VB4IRSTYOS2WB6KNLNJCWU","short_pith_number":"pith:7MY4VB4I","schema_version":"1.0","canonical_sha256":"fb31ca87888ca7874b560f94d5b522b51dbbbbaddf097076d91617830623cae4","source":{"kind":"arxiv","id":"2606.02630","version":1},"attestation_state":"computed","paper":{"title":"MultiTurnPSB: Evaluating Multi-Turn Jailbreak Attacks an dClassifier-Based Defenses for Medical AI Safety","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Anushka Sheoran, Yiduo Hao","submitted_at":"2026-05-30T10:09:53Z","abstract_excerpt":"Patient-facing medical chatbots are commonly evaluated on single-turn prompts, yet real users push back after refusals, add urgency, and invoke authority. We introduce MultiTurnPSB, a four-turn adversarial extension of PatientSafetyBench, and evaluate GPT-4.1-mini under fixed template, template-adaptive, and live adversarial attacks. Unsafe responses rise from 35% to nearly 80% by Turn 4 under live attack. Under the same adversary, GPT-4.1-mini and Claude Sonnet 4.5 are statistically indistinguishable at baseline but diverge to a 19x gap by Turn 4, a difference invisible to single-turn evaluat"},"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":"2606.02630","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-05-30T10:09:53Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"45447795e804bdd5d1a3831e5052738a312f691f042f9f76ca33c5fc7120a538","abstract_canon_sha256":"d2d442d86d2039b9d1b68a1e190089032eb992a334cf44902dbb00b0c6c177c5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-03T00:05:05.134035Z","signature_b64":"Qh8nTT+G7jU+0tVyGc2mzE9OrHx0ADDZtppBmL+Mcz++vhF1JW5M8kMH0RLZXHMUjQ4REQDDmmcMY/ktyCM8BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fb31ca87888ca7874b560f94d5b522b51dbbbbaddf097076d91617830623cae4","last_reissued_at":"2026-06-03T00:05:05.133575Z","signature_status":"signed_v1","first_computed_at":"2026-06-03T00:05:05.133575Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MultiTurnPSB: Evaluating Multi-Turn Jailbreak Attacks an dClassifier-Based Defenses for Medical AI Safety","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Anushka Sheoran, Yiduo Hao","submitted_at":"2026-05-30T10:09:53Z","abstract_excerpt":"Patient-facing medical chatbots are commonly evaluated on single-turn prompts, yet real users push back after refusals, add urgency, and invoke authority. We introduce MultiTurnPSB, a four-turn adversarial extension of PatientSafetyBench, and evaluate GPT-4.1-mini under fixed template, template-adaptive, and live adversarial attacks. Unsafe responses rise from 35% to nearly 80% by Turn 4 under live attack. Under the same adversary, GPT-4.1-mini and Claude Sonnet 4.5 are statistically indistinguishable at baseline but diverge to a 19x gap by Turn 4, a difference invisible to single-turn evaluat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.02630","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/2606.02630/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":"2606.02630","created_at":"2026-06-03T00:05:05.133634+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.02630v1","created_at":"2026-06-03T00:05:05.133634+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.02630","created_at":"2026-06-03T00:05:05.133634+00:00"},{"alias_kind":"pith_short_12","alias_value":"7MY4VB4IRSTY","created_at":"2026-06-03T00:05:05.133634+00:00"},{"alias_kind":"pith_short_16","alias_value":"7MY4VB4IRSTYOS2W","created_at":"2026-06-03T00:05:05.133634+00:00"},{"alias_kind":"pith_short_8","alias_value":"7MY4VB4I","created_at":"2026-06-03T00:05:05.133634+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/7MY4VB4IRSTYOS2WB6KNLNJCWU","json":"https://pith.science/pith/7MY4VB4IRSTYOS2WB6KNLNJCWU.json","graph_json":"https://pith.science/api/pith-number/7MY4VB4IRSTYOS2WB6KNLNJCWU/graph.json","events_json":"https://pith.science/api/pith-number/7MY4VB4IRSTYOS2WB6KNLNJCWU/events.json","paper":"https://pith.science/paper/7MY4VB4I"},"agent_actions":{"view_html":"https://pith.science/pith/7MY4VB4IRSTYOS2WB6KNLNJCWU","download_json":"https://pith.science/pith/7MY4VB4IRSTYOS2WB6KNLNJCWU.json","view_paper":"https://pith.science/paper/7MY4VB4I","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.02630&json=true","fetch_graph":"https://pith.science/api/pith-number/7MY4VB4IRSTYOS2WB6KNLNJCWU/graph.json","fetch_events":"https://pith.science/api/pith-number/7MY4VB4IRSTYOS2WB6KNLNJCWU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7MY4VB4IRSTYOS2WB6KNLNJCWU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7MY4VB4IRSTYOS2WB6KNLNJCWU/action/storage_attestation","attest_author":"https://pith.science/pith/7MY4VB4IRSTYOS2WB6KNLNJCWU/action/author_attestation","sign_citation":"https://pith.science/pith/7MY4VB4IRSTYOS2WB6KNLNJCWU/action/citation_signature","submit_replication":"https://pith.science/pith/7MY4VB4IRSTYOS2WB6KNLNJCWU/action/replication_record"}},"created_at":"2026-06-03T00:05:05.133634+00:00","updated_at":"2026-06-03T00:05:05.133634+00:00"}