{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YJ6ZPBSUGZ6BOS3M3U6M5XTFQB","short_pith_number":"pith:YJ6ZPBSU","schema_version":"1.0","canonical_sha256":"c27d978654367c174b6cdd3ccede65807c27a69154c61b69677e6c8eeb70b343","source":{"kind":"arxiv","id":"2507.21820","version":1},"attestation_state":"computed","paper":{"title":"Anyone Can Jailbreak: Prompt-Based Attacks on LLMs and T2Is","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ahmed B Mustafa, Michael P Pound, Shreyank N Gowda, Yang Lu, Zihan Ye","submitted_at":"2025-07-29T13:55:23Z","abstract_excerpt":"Despite significant advancements in alignment and content moderation, large language models (LLMs) and text-to-image (T2I) systems remain vulnerable to prompt-based attacks known as jailbreaks. Unlike traditional adversarial examples requiring expert knowledge, many of today's jailbreaks are low-effort, high-impact crafted by everyday users with nothing more than cleverly worded prompts. This paper presents a systems-style investigation into how non-experts reliably circumvent safety mechanisms through techniques such as multi-turn narrative escalation, lexical camouflage, implication chaining"},"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":"2507.21820","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-07-29T13:55:23Z","cross_cats_sorted":[],"title_canon_sha256":"c71783fc64efb100c879149b6a69fcd4ffbca54c0a916d3ca83cf04d77b1910d","abstract_canon_sha256":"f77ebda692bbaa2bf663016466d9bb81503c05bc42a2746882d24d4d5943299d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:45:08.160344Z","signature_b64":"u/AtR9rcoshQfiH57hUfhKuy4Jzy6IGRO5HVZlB0+SX+6zCJ3QKZjr/z2TKIl7sBy6yPvx3hiMOYEHDvmiyXCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c27d978654367c174b6cdd3ccede65807c27a69154c61b69677e6c8eeb70b343","last_reissued_at":"2026-07-05T11:45:08.159857Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:45:08.159857Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Anyone Can Jailbreak: Prompt-Based Attacks on LLMs and T2Is","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ahmed B Mustafa, Michael P Pound, Shreyank N Gowda, Yang Lu, Zihan Ye","submitted_at":"2025-07-29T13:55:23Z","abstract_excerpt":"Despite significant advancements in alignment and content moderation, large language models (LLMs) and text-to-image (T2I) systems remain vulnerable to prompt-based attacks known as jailbreaks. Unlike traditional adversarial examples requiring expert knowledge, many of today's jailbreaks are low-effort, high-impact crafted by everyday users with nothing more than cleverly worded prompts. This paper presents a systems-style investigation into how non-experts reliably circumvent safety mechanisms through techniques such as multi-turn narrative escalation, lexical camouflage, implication chaining"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.21820","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/2507.21820/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":"2507.21820","created_at":"2026-07-05T11:45:08.159912+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.21820v1","created_at":"2026-07-05T11:45:08.159912+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.21820","created_at":"2026-07-05T11:45:08.159912+00:00"},{"alias_kind":"pith_short_12","alias_value":"YJ6ZPBSUGZ6B","created_at":"2026-07-05T11:45:08.159912+00:00"},{"alias_kind":"pith_short_16","alias_value":"YJ6ZPBSUGZ6BOS3M","created_at":"2026-07-05T11:45:08.159912+00:00"},{"alias_kind":"pith_short_8","alias_value":"YJ6ZPBSU","created_at":"2026-07-05T11:45:08.159912+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.26999","citing_title":"Prompt Injection Detection is Regime-Dependent: A Deployment-Aware Evaluation with Interpretable Structural Signals","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2604.06954","citing_title":"Compression as an Adversarial Amplifier Through Decision Space Reduction","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18775","citing_title":"An Empirical Study of Multi-Generation Sampling for Jailbreak Detection in Large Language Models","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YJ6ZPBSUGZ6BOS3M3U6M5XTFQB","json":"https://pith.science/pith/YJ6ZPBSUGZ6BOS3M3U6M5XTFQB.json","graph_json":"https://pith.science/api/pith-number/YJ6ZPBSUGZ6BOS3M3U6M5XTFQB/graph.json","events_json":"https://pith.science/api/pith-number/YJ6ZPBSUGZ6BOS3M3U6M5XTFQB/events.json","paper":"https://pith.science/paper/YJ6ZPBSU"},"agent_actions":{"view_html":"https://pith.science/pith/YJ6ZPBSUGZ6BOS3M3U6M5XTFQB","download_json":"https://pith.science/pith/YJ6ZPBSUGZ6BOS3M3U6M5XTFQB.json","view_paper":"https://pith.science/paper/YJ6ZPBSU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.21820&json=true","fetch_graph":"https://pith.science/api/pith-number/YJ6ZPBSUGZ6BOS3M3U6M5XTFQB/graph.json","fetch_events":"https://pith.science/api/pith-number/YJ6ZPBSUGZ6BOS3M3U6M5XTFQB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YJ6ZPBSUGZ6BOS3M3U6M5XTFQB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YJ6ZPBSUGZ6BOS3M3U6M5XTFQB/action/storage_attestation","attest_author":"https://pith.science/pith/YJ6ZPBSUGZ6BOS3M3U6M5XTFQB/action/author_attestation","sign_citation":"https://pith.science/pith/YJ6ZPBSUGZ6BOS3M3U6M5XTFQB/action/citation_signature","submit_replication":"https://pith.science/pith/YJ6ZPBSUGZ6BOS3M3U6M5XTFQB/action/replication_record"}},"created_at":"2026-07-05T11:45:08.159912+00:00","updated_at":"2026-07-05T11:45:08.159912+00:00"}